d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
36597750de
freebsd-skq/sbin/ipfw/ipfw2.c

3900 lines
90 KiB
C
Raw Normal View History

The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
* Copyright (c) 2002-2003 Luigi Rizzo
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* Copyright (c) 1996 Alex Nash, Paul Traina, Poul-Henning Kamp
* Copyright (c) 1994 Ugen J.S.Antsilevich
*
* Idea and grammar partially left from:
* Copyright (c) 1993 Daniel Boulet
*
* Redistribution and use in source forms, with and without modification,
* are permitted provided that this entire comment appears intact.
*
* Redistribution in binary form may occur without any restrictions.
* Obviously, it would be nice if you gave credit where credit is due
* but requiring it would be too onerous.
*
* This software is provided ``AS IS'' without any warranties of any kind.
*
* NEW command line interface for IP firewall facility
*
* $FreeBSD$
*/
remove some useless #include, document why timeconv.h is needed MFC after: 3 days
2009-01-22 23:25:28 +00:00
#include <sys/types.h>
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#include <sys/socket.h>
#include <sys/sockio.h>
#include <sys/sysctl.h>
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
#include "ipfw2.h"
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#include <ctype.h>
#include <err.h>
#include <errno.h>
#include <grp.h>
#include <netdb.h>
#include <pwd.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sysexits.h>
put the altq-related functions into a separate file. Minor cleanup of the includes used by the various source files, including annotations of why certain headers are used.
2009-02-01 16:00:49 +00:00
#include <time.h> /* ctime */
remove some useless #include, document why timeconv.h is needed MFC after: 3 days
2009-01-22 23:25:28 +00:00
#include <timeconv.h> /* _long_to_time */
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
#include <unistd.h>
#include <fcntl.h>
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
#include <net/ethernet.h>
put the altq-related functions into a separate file. Minor cleanup of the includes used by the various source files, including annotations of why certain headers are used.
2009-02-01 16:00:49 +00:00
#include <net/if.h> /* only IFNAMSIZ */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#include <netinet/in.h>
put the altq-related functions into a separate file. Minor cleanup of the includes used by the various source files, including annotations of why certain headers are used.
2009-02-01 16:00:49 +00:00
#include <netinet/in_systm.h> /* only n_short, n_long */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#include <netinet/ip.h>
#include <netinet/ip_icmp.h>
#include <netinet/ip_fw.h>
#include <netinet/tcp.h>
#include <arpa/inet.h>
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
struct cmdline_opts co; /* global options */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
int resvd_set_number = RESVD_SET;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
#define GET_UINT_ARG(arg, min, max, tok, s_x) do { \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (!av[0]) \
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
errx(EX_USAGE, "%s: missing argument", match_value(s_x, tok)); \
if (_substrcmp(*av, "tablearg") == 0) { \
arg = IP_FW_TABLEARG; \
break; \
} \
\
{ \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
long _xval; \
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
char *end; \
\
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
_xval = strtol(*av, &end, 10); \
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
\
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (!isdigit(**av) || *end != '\0' || (_xval == 0 && errno == EINVAL)) \
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
errx(EX_DATAERR, "%s: invalid argument: %s", \
match_value(s_x, tok), *av); \
\
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (errno == ERANGE || _xval < min || _xval > max) \
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
errx(EX_DATAERR, "%s: argument is out of range (%u..%u): %s", \
match_value(s_x, tok), min, max, *av); \
\
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (_xval == IP_FW_TABLEARG) \
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
errx(EX_DATAERR, "%s: illegal argument value: %s", \
match_value(s_x, tok), *av); \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
arg = _xval; \
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
} \
} while (0)
remove a couple of rarely used #define; change PRINT_UINT from a macro to a function (renaming is postponed to reduce clutter)
2009-01-27 07:40:16 +00:00
static void
PRINT_UINT_ARG(const char *str, uint32_t arg)
{
if (str != NULL)
printf("%s",str);
if (arg == IP_FW_TABLEARG)
printf("tablearg");
else
printf("%u", arg);
}
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
static struct _s_x f_tcpflags[] = {
{ "syn", TH_SYN },
{ "fin", TH_FIN },
{ "ack", TH_ACK },
{ "psh", TH_PUSH },
{ "rst", TH_RST },
{ "urg", TH_URG },
{ "tcp flag", 0 },
{ NULL, 0 }
};
static struct _s_x f_tcpopts[] = {
{ "mss", IP_FW_TCPOPT_MSS },
{ "maxseg", IP_FW_TCPOPT_MSS },
{ "window", IP_FW_TCPOPT_WINDOW },
{ "sack", IP_FW_TCPOPT_SACK },
{ "ts", IP_FW_TCPOPT_TS },
{ "timestamp", IP_FW_TCPOPT_TS },
{ "cc", IP_FW_TCPOPT_CC },
{ "tcp option", 0 },
{ NULL, 0 }
};
/*
* IP options span the range 0 to 255 so we need to remap them
* (though in fact only the low 5 bits are significant).
*/
static struct _s_x f_ipopts[] = {
{ "ssrr", IP_FW_IPOPT_SSRR},
{ "lsrr", IP_FW_IPOPT_LSRR},
{ "rr", IP_FW_IPOPT_RR},
{ "ts", IP_FW_IPOPT_TS},
{ "ip option", 0 },
{ NULL, 0 }
};
static struct _s_x f_iptos[] = {
{ "lowdelay", IPTOS_LOWDELAY},
{ "throughput", IPTOS_THROUGHPUT},
{ "reliability", IPTOS_RELIABILITY},
{ "mincost", IPTOS_MINCOST},
Comply with the removal of IPTOS_CE and IPTOS_ECT. Discussed on freebsd-net with no objections. Approved by: njl (mentor), rwatson
2007-10-19 12:48:02 +00:00
{ "congestion", IPTOS_ECN_CE},
{ "ecntransport", IPTOS_ECN_ECT0},
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "ip tos option", 0},
{ NULL, 0 }
};
static struct _s_x limit_masks[] = {
{"all", DYN_SRC_ADDR|DYN_SRC_PORT|DYN_DST_ADDR|DYN_DST_PORT},
{"src-addr", DYN_SRC_ADDR},
{"src-port", DYN_SRC_PORT},
{"dst-addr", DYN_DST_ADDR},
{"dst-port", DYN_DST_PORT},
{NULL, 0}
};
/*
* we use IPPROTO_ETHERTYPE as a fake protocol id to call the print routines
* This is only used in this code.
*/
#define IPPROTO_ETHERTYPE 0x1000
static struct _s_x ether_types[] = {
/*
* Note, we cannot use "-:&/" in the names because they are field
* separators in the type specifications. Also, we use s = NULL as
* end-delimiter, because a type of 0 can be legal.
*/
{ "ip", 0x0800 },
{ "ipv4", 0x0800 },
{ "ipv6", 0x86dd },
{ "arp", 0x0806 },
{ "rarp", 0x8035 },
{ "vlan", 0x8100 },
{ "loop", 0x9000 },
{ "trail", 0x1000 },
{ "at", 0x809b },
{ "atalk", 0x809b },
{ "aarp", 0x80f3 },
{ "pppoe_disc", 0x8863 },
{ "pppoe_sess", 0x8864 },
{ "ipx_8022", 0x00E0 },
{ "ipx_8023", 0x0000 },
{ "ipx_ii", 0x8137 },
{ "ipx_snap", 0x8137 },
{ "ipx", 0x8137 },
{ "ns", 0x0600 },
{ NULL, 0 }
};
Put dummynet-related code in a separate file. To this purpose, add prototypes for global functions in ipfw2.h and move there also the list of tokens used in various places in the code.
2009-01-27 11:06:59 +00:00
static struct _s_x rule_actions[] = {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "accept", TOK_ACCEPT },
{ "pass", TOK_ACCEPT },
{ "allow", TOK_ACCEPT },
{ "permit", TOK_ACCEPT },
{ "count", TOK_COUNT },
{ "pipe", TOK_PIPE },
{ "queue", TOK_QUEUE },
{ "divert", TOK_DIVERT },
{ "tee", TOK_TEE },
Add a ng_ipfw node, implementing a quick and simple interface between ipfw(4) and netgraph(4) facilities. Reviewed by: andre, brooks, julian
2005-02-05 12:06:33 +00:00
{ "netgraph", TOK_NETGRAPH },
{ "ngtee", TOK_NGTEE },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "fwd", TOK_FORWARD },
{ "forward", TOK_FORWARD },
{ "skipto", TOK_SKIPTO },
{ "deny", TOK_DENY },
{ "drop", TOK_DENY },
{ "reject", TOK_REJECT },
* Add dynamic sysctl for net.inet6.ip6.fw. * Correct handling of IPv6 Extension Headers. * Add unreach6 code. * Add logging for IPv6. Submitted by: sysctl handling derived from patch from ume needed for ip6fw Obtained from: is_icmp6_query and send_reject6 derived from similar functions of netinet6,ip6fw Reviewed by: ume, gnn; silence on ipfw@ Test setup provided by: CK Software GmbH MFC after: 6 days
2005-08-13 11:02:34 +00:00
{ "reset6", TOK_RESET6 },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "reset", TOK_RESET },
* Add dynamic sysctl for net.inet6.ip6.fw. * Correct handling of IPv6 Extension Headers. * Add unreach6 code. * Add logging for IPv6. Submitted by: sysctl handling derived from patch from ume needed for ip6fw Obtained from: is_icmp6_query and send_reject6 derived from similar functions of netinet6,ip6fw Reviewed by: ume, gnn; silence on ipfw@ Test setup provided by: CK Software GmbH MFC after: 6 days
2005-08-13 11:02:34 +00:00
{ "unreach6", TOK_UNREACH6 },
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
{ "unreach", TOK_UNREACH },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "check-state", TOK_CHECKSTATE },
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
{ "//", TOK_COMMENT },
Summer of Code 2005: improve libalias - part 2 of 2 With the second (and last) part of my previous Summer of Code work, we get: -ipfw's in kernel nat -redirect_* and LSNAT support General information about nat syntax and some examples are available in the ipfw (8) man page. The redirect and LSNAT syntax are identical to natd, so please refer to natd (8) man page. To enable in kernel nat in rc.conf, two options were added: o firewall_nat_enable: equivalent to natd_enable o firewall_nat_interface: equivalent to natd_interface Remember to set net.inet.ip.fw.one_pass to 0, if you want the packet to continue being checked by the firewall ruleset after being (de)aliased. NOTA BENE: due to some problems with libalias architecture, in kernel nat won't work with TSO enabled nic, thus you have to disable TSO via ifconfig (ifconfig foo0 -tso). Approved by: glebius (mentor)
2006-12-29 21:59:17 +00:00
{ "nat", TOK_NAT },
Implement an ipfw action to reassemble ip packets: reass.
2009-04-01 20:23:47 +00:00
{ "reass", TOK_REASS },
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
{ "setfib", TOK_SETFIB },
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
{ NULL, 0 } /* terminator */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
};
Put dummynet-related code in a separate file. To this purpose, add prototypes for global functions in ipfw2.h and move there also the list of tokens used in various places in the code.
2009-01-27 11:06:59 +00:00
static struct _s_x rule_action_params[] = {
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
{ "altq", TOK_ALTQ },
{ "log", TOK_LOG },
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
{ "tag", TOK_TAG },
{ "untag", TOK_UNTAG },
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
{ NULL, 0 } /* terminator */
};
implement a new match option, lookup {dst-ip|src-ip|dst-port|src-port|uid|jail} N which searches the specified field in table N and sets tablearg accordingly. With dst-ip or src-ip the option replicates two existing options. When used with other arguments, the option can be useful to quickly dispatch traffic based on other fields. Work supported by the Onelab project. MFC after: 1 week
2009-12-15 09:46:27 +00:00
/*
* The 'lookup' instruction accepts one of the following arguments.
* -1 is a terminator for the list.
* Arguments are passed as v[1] in O_DST_LOOKUP options.
*/
static int lookup_key[] = {
TOK_DSTIP, TOK_SRCIP, TOK_DSTPORT, TOK_SRCPORT,
Implement "lookup dscp N" which does a lookup of the DSCP (top 6 bits of ip->ip_tos) in a table. This can be useful to direct traffic to different pipes/queues according to the DSCP of the packet, as follows: ipfw add 100 queue tablearg lookup dscp 3 // table 3 maps dscp->queue This change is a no-op (but harmless) until the two-line kernel side is committed, which will happen shortly.
2010-03-15 15:43:35 +00:00
TOK_UID, TOK_JAIL, TOK_DSCP, -1 };
implement a new match option, lookup {dst-ip|src-ip|dst-port|src-port|uid|jail} N which searches the specified field in table N and sets tablearg accordingly. With dst-ip or src-ip the option replicates two existing options. When used with other arguments, the option can be useful to quickly dispatch traffic based on other fields. Work supported by the Onelab project. MFC after: 1 week
2009-12-15 09:46:27 +00:00
Put dummynet-related code in a separate file. To this purpose, add prototypes for global functions in ipfw2.h and move there also the list of tokens used in various places in the code.
2009-01-27 11:06:59 +00:00
static struct _s_x rule_options[] = {
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
{ "tagged", TOK_TAGGED },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "uid", TOK_UID },
{ "gid", TOK_GID },
Add the ability to associate ipfw rules with a specific prison ID. Since the only thing truly unique about a prison is it's ID, I figured this would be the most granular way of handling this. This commit makes the following changes: - Adds tokenizing and parsing for the ``jail'' command line option to the ipfw(8) userspace utility. - Append the ipfw opcode list with O_JAIL. - While Iam here, add a comment informing others that if they want to add additional opcodes, they should append them to the end of the list to avoid ABI breakage. - Add ``fw_prid'' to the ipfw ucred cache structure. - When initializing ucred cache, if the process is jailed, set fw_prid to the prison ID, otherwise set it to -1. - Update man page to reflect these changes. This change was a strong motivator behind the ucred caching mechanism in ipfw. A sample usage of this new functionality could be: ipfw add count ip from any to any jail 2 It should be noted that because ucred based constraints are only implemented for TCP and UDP packets, the same applies for jail associations. Conceptual head nod by: pjd Reviewed by: rwatson Approved by: bmilekic (mentor)
2004-08-12 22:06:55 +00:00
{ "jail", TOK_JAIL },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "in", TOK_IN },
{ "limit", TOK_LIMIT },
{ "keep-state", TOK_KEEPSTATE },
{ "bridged", TOK_LAYER2 },
{ "layer2", TOK_LAYER2 },
{ "out", TOK_OUT },
Add support to IPFW for classification based on "diverted" status (that is, input via a divert socket).
2004-10-03 00:26:35 +00:00
{ "diverted", TOK_DIVERTED },
{ "diverted-loopback", TOK_DIVERTEDLOOPBACK },
{ "diverted-output", TOK_DIVERTEDOUTPUT },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "xmit", TOK_XMIT },
{ "recv", TOK_RECV },
{ "via", TOK_VIA },
{ "fragment", TOK_FRAG },
{ "frag", TOK_FRAG },
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
{ "fib", TOK_FIB },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "ipoptions", TOK_IPOPTS },
{ "ipopts", TOK_IPOPTS },
{ "iplen", TOK_IPLEN },
{ "ipid", TOK_IPID },
{ "ipprecedence", TOK_IPPRECEDENCE },
Implement "lookup dscp N" which does a lookup of the DSCP (top 6 bits of ip->ip_tos) in a table. This can be useful to direct traffic to different pipes/queues according to the DSCP of the packet, as follows: ipfw add 100 queue tablearg lookup dscp 3 // table 3 maps dscp->queue This change is a no-op (but harmless) until the two-line kernel side is committed, which will happen shortly.
2010-03-15 15:43:35 +00:00
{ "dscp", TOK_DSCP },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "iptos", TOK_IPTOS },
{ "ipttl", TOK_IPTTL },
{ "ipversion", TOK_IPVER },
{ "ipver", TOK_IPVER },
{ "estab", TOK_ESTAB },
{ "established", TOK_ESTAB },
{ "setup", TOK_SETUP },
Add support to IPFW for matching by TCP data length.
2004-10-03 00:47:15 +00:00
{ "tcpdatalen", TOK_TCPDATALEN },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "tcpflags", TOK_TCPFLAGS },
{ "tcpflgs", TOK_TCPFLAGS },
{ "tcpoptions", TOK_TCPOPTS },
{ "tcpopts", TOK_TCPOPTS },
{ "tcpseq", TOK_TCPSEQ },
{ "tcpack", TOK_TCPACK },
{ "tcpwin", TOK_TCPWIN },
A bunch of minor fixes: * accept "icmptype" as an alias for "icmptypes"; * remove an extra whitespace after "log" rules; * print correctly the "limit" masks; * correct a typo in parsing dummynet arguments (this caused a coredump); * do not allow specifying both "check-state" and "limit", they are (and have always been) mutually exclusive; * remove an extra print of the rule before installing it; * make stdout buffered -- otherwise, if you log its output with syslog, you will see one entry for each printf(). Rather unpleasant.
2002-07-13 15:57:23 +00:00
{ "icmptype", TOK_ICMPTYPES },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "icmptypes", TOK_ICMPTYPES },
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
{ "dst-ip", TOK_DSTIP },
{ "src-ip", TOK_SRCIP },
{ "dst-port", TOK_DSTPORT },
{ "src-port", TOK_SRCPORT },
{ "proto", TOK_PROTO },
{ "MAC", TOK_MAC },
{ "mac", TOK_MAC },
{ "mac-type", TOK_MACTYPE },
Add a 'verrevpath' option that verifies the interface that a packet comes in on is the same interface that we would route out of to get to the packet's source address. Essentially automates an anti-spoofing check using the information in the routing table. Experimental. The usage and rule format for the feature may still be subject to change.
2003-03-15 01:13:00 +00:00
{ "verrevpath", TOK_VERREVPATH },
Add the option versrcreach to verify that a valid route to the source address of a packet exists in the routing table. The default route is ignored because it would match everything and render the check pointless. This option is very useful for routers with a complete view of the Internet (BGP) in the routing table to reject packets with spoofed or unrouteable source addresses. Example: ipfw add 1000 deny ip from any to any not versrcreach also known in Cisco-speak as: ip verify unicast source reachable-via any Reviewed by: luigi
2004-04-23 14:28:38 +00:00
{ "versrcreach", TOK_VERSRCREACH },
New ipfw option "antispoof": For incoming packets, the packet's source address is checked if it belongs to a directly connected network. If the network is directly connected, then the interface the packet came on in is compared to the interface the network is connected to. When incoming interface and directly connected interface are not the same, the packet does not match. Usage example: ipfw add deny ip from any to any not antispoof in Manpage education by: ru
2004-08-09 16:12:10 +00:00
{ "antispoof", TOK_ANTISPOOF },
Implement the 'ipsec' option to match packets coming out of an ipsec tunnel. Should work with both regular and fast ipsec (mutually exclusive). See manpage for more details. Submitted by: Ari Suutari (ari.suutari@syncrontech.com) Revised by: sam MFC after: 1 week
2003-07-04 21:42:32 +00:00
{ "ipsec", TOK_IPSEC },
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
{ "icmp6type", TOK_ICMP6TYPES },
{ "icmp6types", TOK_ICMP6TYPES },
{ "ext6hdr", TOK_EXT6HDR},
{ "flow-id", TOK_FLOWID},
{ "ipv6", TOK_IPV6},
{ "ip6", TOK_IPV6},
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
{ "ipv4", TOK_IPV4},
{ "ip4", TOK_IPV4},
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
{ "dst-ipv6", TOK_DSTIP6},
{ "dst-ip6", TOK_DSTIP6},
{ "src-ipv6", TOK_SRCIP6},
{ "src-ip6", TOK_SRCIP6},
implement a new match option, lookup {dst-ip|src-ip|dst-port|src-port|uid|jail} N which searches the specified field in table N and sets tablearg accordingly. With dst-ip or src-ip the option replicates two existing options. When used with other arguments, the option can be useful to quickly dispatch traffic based on other fields. Work supported by the Onelab project. MFC after: 1 week
2009-12-15 09:46:27 +00:00
{ "lookup", TOK_LOOKUP},
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
{ "//", TOK_COMMENT },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "not", TOK_NOT }, /* pseudo option */
{ "!", /* escape ? */ TOK_NOT }, /* pseudo option */
{ "or", TOK_OR }, /* pseudo option */
{ "|", /* escape */ TOK_OR }, /* pseudo option */
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
{ "{", TOK_STARTBRACE }, /* pseudo option */
{ "(", TOK_STARTBRACE }, /* pseudo option */
{ "}", TOK_ENDBRACE }, /* pseudo option */
{ ")", TOK_ENDBRACE }, /* pseudo option */
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
{ NULL, 0 } /* terminator */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
};
fix printing of uint64_t values, so we can use WARNS=2
2009-01-27 20:26:45 +00:00
/*
* The following is used to generate a printable argument for
* 64-bit numbers, irrespective of platform alignment and bit size.
* Because all the printf in this program use %llu as a format,
* we just return an unsigned long long, which is larger than
* we need in certain cases, but saves the hassle of using
* PRIu64 as a format specifier.
* We don't care about inlining, this is not performance critical code.
*/
unsigned long long
align_uint64(const uint64_t *pll)
{
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint64_t ret;
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
bcopy (pll, &ret, sizeof(ret));
return ret;
Remove spurious semicolons. Approved by: das (mentor) Reviewed by: ipfw@
2004-05-18 13:57:37 +00:00
}
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
void *
wrap all malloc/calloc/realloc calls so they exit on failure without having to check in each place. Remove an wrong strdup from previous commit.
2009-01-26 14:26:35 +00:00
safe_calloc(size_t number, size_t size)
{
void *ret = calloc(number, size);
if (ret == NULL)
err(EX_OSERR, "calloc");
return ret;
}
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
void *
wrap all malloc/calloc/realloc calls so they exit on failure without having to check in each place. Remove an wrong strdup from previous commit.
2009-01-26 14:26:35 +00:00
safe_realloc(void *ptr, size_t size)
{
void *ret = realloc(ptr, size);
if (ret == NULL)
err(EX_OSERR, "realloc");
return ret;
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
/*
* conditionally runs the command.
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
* Selected options or negative -> getsockopt
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
*/
Put dummynet-related code in a separate file. To this purpose, add prototypes for global functions in ipfw2.h and move there also the list of tokens used in various places in the code.
2009-01-27 11:06:59 +00:00
int
Apply a bandaid to get this working on sparc64 again; the introduction of do_cmd() broke things, because this function assumes that a socklen_t is large enough to hold a pointer. A real solution to this problem would be a rewrite of do_cmd() to treat the optlen parameter consistently and not use it to carry a pointer or integer dependent on the context.
2003-09-04 15:57:37 +00:00
do_cmd(int optname, void *optval, uintptr_t optlen)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
{
static int s = -1; /* the socket */
int i;
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.test_only)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
return 0;
if (s == -1)
s = socket(AF_INET, SOCK_RAW, IPPROTO_RAW);
if (s < 0)
err(EX_UNAVAILABLE, "socket");
if (optname == IP_FW_GET || optname == IP_DUMMYNET_GET ||
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
optname == IP_FW_ADD || optname == IP_FW_TABLE_LIST ||
Summer of Code 2005: improve libalias - part 2 of 2 With the second (and last) part of my previous Summer of Code work, we get: -ipfw's in kernel nat -redirect_* and LSNAT support General information about nat syntax and some examples are available in the ipfw (8) man page. The redirect and LSNAT syntax are identical to natd, so please refer to natd (8) man page. To enable in kernel nat in rc.conf, two options were added: o firewall_nat_enable: equivalent to natd_enable o firewall_nat_interface: equivalent to natd_interface Remember to set net.inet.ip.fw.one_pass to 0, if you want the packet to continue being checked by the firewall ruleset after being (de)aliased. NOTA BENE: due to some problems with libalias architecture, in kernel nat won't work with TSO enabled nic, thus you have to disable TSO via ifconfig (ifconfig foo0 -tso). Approved by: glebius (mentor)
2006-12-29 21:59:17 +00:00
optname == IP_FW_TABLE_GETSIZE ||
optname == IP_FW_NAT_GET_CONFIG ||
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
optname < 0 ||
optname == IP_FW_NAT_GET_LOG) {
if (optname < 0)
optname = -optname;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
i = getsockopt(s, IPPROTO_IP, optname, optval,
(socklen_t *)optlen);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
} else {
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
i = setsockopt(s, IPPROTO_IP, optname, optval, optlen);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
return i;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/**
* match_token takes a table and a string, returns the value associated
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
* with the string (-1 in case of failure).
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
Put dummynet-related code in a separate file. To this purpose, add prototypes for global functions in ipfw2.h and move there also the list of tokens used in various places in the code.
2009-01-27 11:06:59 +00:00
int
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
match_token(struct _s_x *table, char *string)
{
struct _s_x *pt;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
uint i = strlen(string);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
for (pt = table ; i && pt->s != NULL ; pt++)
if (strlen(pt->s) == i && !bcmp(string, pt->s, i))
return pt->x;
return -1;
Remove spurious semicolons. Approved by: das (mentor) Reviewed by: ipfw@
2004-05-18 13:57:37 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
/**
* match_value takes a table and a value, returns the string associated
* with the value (NULL in case of failure).
*/
Put nat and ipv6 support in their own files. Usual moving of code with no changes from ipfw2.c to the newly created files, and addition of prototypes to ipfw2.h I have added forward declarations for ipfw_insn_* in ipfw2.h to avoid a global dependency on ip_fw.h
2009-01-27 12:01:30 +00:00
char const *
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
match_value(struct _s_x *p, int value)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
for (; p->s != NULL; p++)
if (p->x == value)
return p->s;
return NULL;
}
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
/*
* _substrcmp takes two strings and returns 1 if they do not match,
* and 0 if they match exactly or the first string is a sub-string
* of the second. A warning is printed to stderr in the case that the
* first string is a sub-string of the second.
*
* This function will be removed in the future through the usual
* deprecation process.
*/
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
int
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
_substrcmp(const char *str1, const char* str2)
{
if (strncmp(str1, str2, strlen(str1)) != 0)
return 1;
if (strlen(str1) != strlen(str2))
warnx("DEPRECATED: '%s' matched '%s' as a sub-string",
str1, str2);
return 0;
}
/*
* _substrcmp2 takes three strings and returns 1 if the first two do not match,
* and 0 if they match exactly or the second string is a sub-string
* of the first. A warning is printed to stderr in the case that the
* first string does not match the third.
*
* This function exists to warn about the bizzare construction
* strncmp(str, "by", 2) which is used to allow people to use a shotcut
* for "bytes". The problem is that in addition to accepting "by",
* "byt", "byte", and "bytes", it also excepts "by_rabid_dogs" and any
* other string beginning with "by".
*
* This function will be removed in the future through the usual
* deprecation process.
*/
Put dummynet-related code in a separate file. To this purpose, add prototypes for global functions in ipfw2.h and move there also the list of tokens used in various places in the code.
2009-01-27 11:06:59 +00:00
int
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
_substrcmp2(const char *str1, const char* str2, const char* str3)
{
if (strncmp(str1, str2, strlen(str2)) != 0)
return 1;
if (strcmp(str1, str3) != 0)
warnx("DEPRECATED: '%s' matched '%s'",
str1, str3);
return 0;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* prints one port, symbolic or numeric
*/
static void
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
print_port(int proto, uint16_t port)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
if (proto == IPPROTO_ETHERTYPE) {
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
char const *s;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.do_resolv && (s = match_value(ether_types, port)) )
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
printf("%s", s);
else
printf("0x%04x", port);
} else {
struct servent *se = NULL;
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.do_resolv) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
struct protoent *pe = getprotobynumber(proto);
se = getservbyport(htons(port), pe ? pe->p_name : NULL);
}
if (se)
printf("%s", se->s_name);
else
printf("%d", port);
}
}
Put dummynet-related code in a separate file. To this purpose, add prototypes for global functions in ipfw2.h and move there also the list of tokens used in various places in the code.
2009-01-27 11:06:59 +00:00
static struct _s_x _port_name[] = {
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
{"dst-port", O_IP_DSTPORT},
{"src-port", O_IP_SRCPORT},
{"ipid", O_IPID},
{"iplen", O_IPLEN},
{"ipttl", O_IPTTL},
{"mac-type", O_MAC_TYPE},
Add support to IPFW for matching by TCP data length.
2004-10-03 00:47:15 +00:00
{"tcpdatalen", O_TCPDATALEN},
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
{"tagged", O_TAGGED},
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
{NULL, 0}
};
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
* Print the values in a list 16-bit items of the types above.
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* XXX todo: add support for mask.
*/
static void
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
print_newports(ipfw_insn_u16 *cmd, int proto, int opcode)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint16_t *p = cmd->ports;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int i;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
char const *sep;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Add support for multiple values and ranges for the "iplen", "ipttl", "ipid" options. This feature has been requested by several users. On passing, fix some minor bugs in the parser. This change is fully backward compatible so if you have an old /sbin/ipfw and a new kernel you are not in trouble (but you need to update /sbin/ipfw if you want to use the new features). Document the changes in the manpage. Now you can write things like ipfw add skipto 1000 iplen 0-500 which some people were asking to give preferential treatment to short packets. The 'MFC after' is just set as a reminder, because I still need to merge the Alpha/Sparc64 fixes for ipfw2 (which unfortunately change the size of certain kernel structures; not that it matters a lot since ipfw2 is entirely optional and not the default...) PR: bin/48015 MFC after: 1 week
2003-06-22 17:33:19 +00:00
if (opcode != 0) {
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
sep = match_value(_port_name, opcode);
if (sep == NULL)
Add support for multiple values and ranges for the "iplen", "ipttl", "ipid" options. This feature has been requested by several users. On passing, fix some minor bugs in the parser. This change is fully backward compatible so if you have an old /sbin/ipfw and a new kernel you are not in trouble (but you need to update /sbin/ipfw if you want to use the new features). Document the changes in the manpage. Now you can write things like ipfw add skipto 1000 iplen 0-500 which some people were asking to give preferential treatment to short packets. The 'MFC after' is just set as a reminder, because I still need to merge the Alpha/Sparc64 fixes for ipfw2 (which unfortunately change the size of certain kernel structures; not that it matters a lot since ipfw2 is entirely optional and not the default...) PR: bin/48015 MFC after: 1 week
2003-06-22 17:33:19 +00:00
sep = "???";
printf (" %s", sep);
}
sep = " ";
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
for (i = F_LEN((ipfw_insn *)cmd) - 1; i > 0; i--, p += 2) {
add a missing format in a printf Detected building with gcc 4.3.3 MFC after: 3 days
2009-06-08 10:53:18 +00:00
printf("%s", sep);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
print_port(proto, p[0]);
if (p[0] != p[1]) {
printf("-");
print_port(proto, p[1]);
}
sep = ",";
}
}
/*
* Like strtol, but also translates service names into port numbers
* for some protocols.
* In particular:
* proto == -1 disables the protocol check;
* proto == IPPROTO_ETHERTYPE looks up an internal table
* proto == <some value in /etc/protocols> matches the values there.
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
* Returns *end == s in case the parameter is not found.
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
static int
strtoport(char *s, char **end, int base, int proto)
{
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
char *p, *buf;
char *s1;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int i;
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
*end = s; /* default - not found */
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
if (*s == '\0')
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
return 0; /* not found */
Kill EOL whitespaces, style(9) fix.
2002-11-06 15:09:34 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (isdigit(*s))
return strtol(s, end, base);
/*
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
* find separator. '\\' escapes the next char.
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
for (s1 = s; *s1 && (isalnum(*s1) || *s1 == '\\') ; s1++)
if (*s1 == '\\' && s1[1] != '\0')
s1++;
wrap all malloc/calloc/realloc calls so they exit on failure without having to check in each place. Remove an wrong strdup from previous commit.
2009-01-26 14:26:35 +00:00
buf = safe_calloc(s1 - s + 1, 1);
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
/*
* copy into a buffer skipping backslashes
*/
for (p = s, i = 0; p != s1 ; p++)
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
if (*p != '\\')
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
buf[i++] = *p;
buf[i++] = '\0';
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (proto == IPPROTO_ETHERTYPE) {
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
i = match_token(ether_types, buf);
free(buf);
if (i != -1) { /* found */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*end = s1;
return i;
}
} else {
struct protoent *pe = NULL;
struct servent *se;
if (proto != 0)
pe = getprotobynumber(proto);
setservent(1);
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
se = getservbyname(buf, pe ? pe->p_name : NULL);
free(buf);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (se != NULL) {
*end = s1;
return ntohs(se->s_port);
}
}
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
return 0; /* not found */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
/*
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
* Fill the body of the command with the list of port ranges.
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
static int
fill_newports(ipfw_insn_u16 *cmd, char *av, int proto)
{
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint16_t a, b, *p = cmd->ports;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int i = 0;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
char *s = av;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
while (*s) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
a = strtoport(av, &s, 0, proto);
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
if (s == av) /* empty or invalid argument */
return (0);
switch (*s) {
case '-': /* a range */
av = s + 1;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
b = strtoport(av, &s, 0, proto);
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
/* Reject expressions like '1-abc' or '1-2-3'. */
if (s == av || (*s != ',' && *s != '\0'))
return (0);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
p[0] = a;
p[1] = b;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
break;
case ',': /* comma separated list */
case '\0':
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
p[0] = p[1] = a;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
break;
default:
warnx("port list: invalid separator <%c> in <%s>",
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
*s, av);
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
return (0);
}
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
i++;
p += 2;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
av = s + 1;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
if (i > 0) {
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
if (i + 1 > F_LEN_MASK)
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
errx(EX_DATAERR, "too many ports/ranges\n");
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
cmd->o.len |= i + 1; /* leave F_NOT and F_OR untouched */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
return (i);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
static struct _s_x icmpcodes[] = {
{ "net", ICMP_UNREACH_NET },
{ "host", ICMP_UNREACH_HOST },
{ "protocol", ICMP_UNREACH_PROTOCOL },
{ "port", ICMP_UNREACH_PORT },
{ "needfrag", ICMP_UNREACH_NEEDFRAG },
{ "srcfail", ICMP_UNREACH_SRCFAIL },
{ "net-unknown", ICMP_UNREACH_NET_UNKNOWN },
{ "host-unknown", ICMP_UNREACH_HOST_UNKNOWN },
{ "isolated", ICMP_UNREACH_ISOLATED },
{ "net-prohib", ICMP_UNREACH_NET_PROHIB },
{ "host-prohib", ICMP_UNREACH_HOST_PROHIB },
{ "tosnet", ICMP_UNREACH_TOSNET },
{ "toshost", ICMP_UNREACH_TOSHOST },
{ "filter-prohib", ICMP_UNREACH_FILTER_PROHIB },
{ "host-precedence", ICMP_UNREACH_HOST_PRECEDENCE },
{ "precedence-cutoff", ICMP_UNREACH_PRECEDENCE_CUTOFF },
{ NULL, 0 }
};
static void
fill_reject_code(u_short *codep, char *str)
{
int val;
char *s;
val = strtoul(str, &s, 0);
if (s == str || *s != '\0' || val >= 0x100)
val = match_token(icmpcodes, str);
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
if (val < 0)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
errx(EX_DATAERR, "unknown ICMP unreachable code ``%s''", str);
*codep = val;
return;
}
static void
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
print_reject_code(uint16_t code)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
char const *s = match_value(icmpcodes, code);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (s != NULL)
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
printf("unreach %s", s);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
else
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
printf("unreach %u", code);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
/*
* Returns the number of bits set (from left) in a contiguous bitmask,
* or -1 if the mask is not contiguous.
* XXX this needs a proper fix.
* This effectively works on masks in big-endian (network) format.
* when compiled on little endian architectures.
*
* First bit is bit 7 of the first byte -- note, for MAC addresses,
* the first bit on the wire is bit 0 of the first byte.
* len is the max length in bits.
*/
Put nat and ipv6 support in their own files. Usual moving of code with no changes from ipfw2.c to the newly created files, and addition of prototypes to ipfw2.h I have added forward declarations for ipfw_insn_* in ipfw2.h to avoid a global dependency on ip_fw.h
2009-01-27 12:01:30 +00:00
int
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
contigmask(uint8_t *p, int len)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
int i, n;
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
for (i=0; i<len ; i++)
if ( (p[i/8] & (1 << (7 - (i%8)))) == 0) /* first bit unset */
break;
for (n=i+1; n < len; n++)
if ( (p[n/8] & (1 << (7 - (n%8)))) != 0)
return -1; /* mask not contiguous */
return i;
}
/*
* print flags set/clear in the two bitmasks passed as parameters.
* There is a specialized check for f_tcpflags.
*/
static void
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
print_flags(char const *name, ipfw_insn *cmd, struct _s_x *list)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
char const *comma = "";
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int i;
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
uint8_t set = cmd->arg1 & 0xff;
uint8_t clear = (cmd->arg1 >> 8) & 0xff;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (list == f_tcpflags && set == TH_SYN && clear == TH_ACK) {
printf(" setup");
return;
}
printf(" %s ", name);
for (i=0; list[i].x != 0; i++) {
if (set & list[i].x) {
set &= ~list[i].x;
printf("%s%s", comma, list[i].s);
comma = ",";
}
if (clear & list[i].x) {
clear &= ~list[i].x;
printf("%s!%s", comma, list[i].s);
comma = ",";
}
}
}
/*
* Print the ip address contained in a command.
*/
static void
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
print_ip(ipfw_insn_ip *cmd, char const *s)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
struct hostent *he = NULL;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
uint32_t len = F_LEN((ipfw_insn *)cmd);
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint32_t *a = ((ipfw_insn_u32 *)cmd)->d;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
implement a new match option, lookup {dst-ip|src-ip|dst-port|src-port|uid|jail} N which searches the specified field in table N and sets tablearg accordingly. With dst-ip or src-ip the option replicates two existing options. When used with other arguments, the option can be useful to quickly dispatch traffic based on other fields. Work supported by the Onelab project. MFC after: 1 week
2009-12-15 09:46:27 +00:00
if (cmd->o.opcode == O_IP_DST_LOOKUP && len > F_INSN_SIZE(ipfw_insn_u32)) {
uint32_t d = a[1];
const char *arg = "<invalid>";
if (d < sizeof(lookup_key)/sizeof(lookup_key[0]))
arg = match_value(rule_options, lookup_key[d]);
printf("%s lookup %s %d", cmd->o.len & F_NOT ? " not": "",
arg, cmd->o.arg1);
return;
}
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
printf("%s%s ", cmd->o.len & F_NOT ? " not": "", s);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (cmd->o.opcode == O_IP_SRC_ME || cmd->o.opcode == O_IP_DST_ME) {
printf("me");
return;
}
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
if (cmd->o.opcode == O_IP_SRC_LOOKUP ||
cmd->o.opcode == O_IP_DST_LOOKUP) {
printf("table(%u", ((ipfw_insn *)cmd)->arg1);
if (len == F_INSN_SIZE(ipfw_insn_u32))
printf(",%u", *a);
printf(")");
return;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (cmd->o.opcode == O_IP_SRC_SET || cmd->o.opcode == O_IP_DST_SET) {
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint32_t x, *map = (uint32_t *)&(cmd->mask);
syntactic sugar: support range notation such as 1.2.3.4/24{5,6,7,10-20,60-90} for set of ip addresses. Previously you needed to specify every address in the range, which was unconvenient and lead to very long lines. Internally the set is still stored in the same way, just the input and output routines are modified. Manpage update still missing. Perhaps a similar preprocessing step would be useful for port ranges. MFC after: 3 days
2003-06-23 08:20:28 +00:00
int i, j;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
char comma = '{';
x = cmd->o.arg1 - 1;
x = htonl( ~x );
cmd->addr.s_addr = htonl(cmd->addr.s_addr);
printf("%s/%d", inet_ntoa(cmd->addr),
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
contigmask((uint8_t *)&x, 32));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
x = cmd->addr.s_addr = htonl(cmd->addr.s_addr);
x &= 0xff; /* base */
syntactic sugar: support range notation such as 1.2.3.4/24{5,6,7,10-20,60-90} for set of ip addresses. Previously you needed to specify every address in the range, which was unconvenient and lead to very long lines. Internally the set is still stored in the same way, just the input and output routines are modified. Manpage update still missing. Perhaps a similar preprocessing step would be useful for port ranges. MFC after: 3 days
2003-06-23 08:20:28 +00:00
/*
* Print bits and ranges.
* Locate first bit set (i), then locate first bit unset (j).
* If we have 3+ consecutive bits set, then print them as a
* range, otherwise only print the initial bit and rescan.
*/
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
for (i=0; i < cmd->o.arg1; i++)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
if (map[i/32] & (1<<(i & 31))) {
syntactic sugar: support range notation such as 1.2.3.4/24{5,6,7,10-20,60-90} for set of ip addresses. Previously you needed to specify every address in the range, which was unconvenient and lead to very long lines. Internally the set is still stored in the same way, just the input and output routines are modified. Manpage update still missing. Perhaps a similar preprocessing step would be useful for port ranges. MFC after: 3 days
2003-06-23 08:20:28 +00:00
for (j=i+1; j < cmd->o.arg1; j++)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
if (!(map[ j/32] & (1<<(j & 31))))
syntactic sugar: support range notation such as 1.2.3.4/24{5,6,7,10-20,60-90} for set of ip addresses. Previously you needed to specify every address in the range, which was unconvenient and lead to very long lines. Internally the set is still stored in the same way, just the input and output routines are modified. Manpage update still missing. Perhaps a similar preprocessing step would be useful for port ranges. MFC after: 3 days
2003-06-23 08:20:28 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
printf("%c%d", comma, i+x);
syntactic sugar: support range notation such as 1.2.3.4/24{5,6,7,10-20,60-90} for set of ip addresses. Previously you needed to specify every address in the range, which was unconvenient and lead to very long lines. Internally the set is still stored in the same way, just the input and output routines are modified. Manpage update still missing. Perhaps a similar preprocessing step would be useful for port ranges. MFC after: 3 days
2003-06-23 08:20:28 +00:00
if (j>i+2) { /* range has at least 3 elements */
printf("-%d", j-1+x);
i = j-1;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
comma = ',';
}
printf("}");
return;
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
/*
* len == 2 indicates a single IP, whereas lists of 1 or more
* addr/mask pairs have len = (2n+1). We convert len to n so we
* use that to count the number of entries.
*/
for (len = len / 2; len > 0; len--, a += 2) {
int mb = /* mask length */
(cmd->o.opcode == O_IP_SRC || cmd->o.opcode == O_IP_DST) ?
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
32 : contigmask((uint8_t *)&(a[1]), 32);
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (mb == 32 && co.do_resolv)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
he = gethostbyaddr((char *)&(a[0]), sizeof(u_long), AF_INET);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (he != NULL) /* resolved to name */
printf("%s", he->h_name);
else if (mb == 0) /* any */
printf("any");
else { /* numeric IP followed by some kind of mask */
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
printf("%s", inet_ntoa( *((struct in_addr *)&a[0]) ) );
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (mb < 0)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
printf(":%s", inet_ntoa( *((struct in_addr *)&a[1]) ) );
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
else if (mb < 32)
printf("/%d", mb);
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
if (len > 1)
printf(",");
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
/*
* prints a MAC address/mask pair
*/
static void
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
print_mac(uint8_t *addr, uint8_t *mask)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
int l = contigmask(mask, 48);
if (l == 0)
printf(" any");
else {
printf(" %02x:%02x:%02x:%02x:%02x:%02x",
addr[0], addr[1], addr[2], addr[3], addr[4], addr[5]);
if (l == -1)
printf("&%02x:%02x:%02x:%02x:%02x:%02x",
mask[0], mask[1], mask[2],
mask[3], mask[4], mask[5]);
else if (l < 48)
printf("/%d", l);
}
}
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
static void
fill_icmptypes(ipfw_insn_u32 *cmd, char *av)
{
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint8_t type;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
cmd->d[0] = 0;
while (*av) {
if (*av == ',')
av++;
type = strtoul(av, &av, 0);
if (*av != ',' && *av != '\0')
errx(EX_DATAERR, "invalid ICMP type");
if (type > 31)
errx(EX_DATAERR, "ICMP type out of range");
cmd->d[0] |= 1 << type;
}
cmd->o.opcode = O_ICMPTYPE;
cmd->o.len |= F_INSN_SIZE(ipfw_insn_u32);
}
static void
print_icmptypes(ipfw_insn_u32 *cmd)
{
int i;
char sep= ' ';
printf(" icmptypes");
for (i = 0; i < 32; i++) {
if ( (cmd->d[0] & (1 << (i))) == 0)
continue;
printf("%c%d", sep, i);
sep = ',';
}
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* show_ipfw() prints the body of an ipfw rule.
* Because the standard rule has at least proto src_ip dst_ip, we use
* a helper function to produce these entries if not provided explicitly.
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
* The first argument is the list of fields we have, the second is
* the list of fields we want to be printed.
*
* Special cases if we have provided a MAC header:
* + if the rule does not contain IP addresses/ports, do not print them;
* + if the rule does not contain an IP proto, print "all" instead of "ip";
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
*
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
* Once we have 'have_options', IP header fields are printed as options.
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
#define HAVE_PROTO 0x0001
#define HAVE_SRCIP 0x0002
#define HAVE_DSTIP 0x0004
o Make ipfw(8) show rules with mac/mac-type options correctly. Before: $ ipfw -n add 100 count icmp from any to any mac-type 0x01 00100 count icmp 0x0001 $ ipfw -n add 100 count icmp from any to any mac any any 00100 count icmp MAC any any any After: $ ipfw -n add 100 count icmp from any to any mac-type 0x01 00100 count icmp from any to any mac-type 0x0001 $ ipfw -n add 100 count icmp from any to any mac any any 00100 count icmp from any to any MAC any any PR: bin/112244 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-04-30 17:39:30 +00:00
#define HAVE_PROTO4 0x0008
#define HAVE_PROTO6 0x0010
print correctly commands of the form ipfw add 100 allow ip from { 1.2.3.4 or 5.6.7.8 } (note that the above example could be better written as ipfw add 100 allow dst-ip 1.2.3.4,5.6.7.8 Submitted by: Riccardo Panicucci
2010-03-15 18:20:51 +00:00
#define HAVE_IP 0x0100
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
#define HAVE_OPTIONS 0x8000
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
static void
Fix a number of (innocuous) warnings, and remove a useless test. There are still several signed/unsigned warnings left, which require a bit more study for a proper fix. This file has grown beyond reasonable limits. We really need to split it into separate components (ipv4, ipv6, dummynet, nat, table, userland-kernel communication ...) so we can make mainteinance easier. MFC after: 1 weeks
2009-01-20 18:16:31 +00:00
show_prerequisites(int *flags, int want, int cmd __unused)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.comment_only)
Add a -b flag to /sbin/ipfw to print only action and comment for each rule, thus omitting the entire body. This makes the output a lot more readable for complex rulesets (provided, of course, you have annotated your ruleset appropriately!) MFC after: 3 days
2003-12-12 16:14:28 +00:00
return;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
if ( (*flags & HAVE_IP) == HAVE_IP)
*flags |= HAVE_OPTIONS;
if ( !(*flags & HAVE_OPTIONS)) {
Fix a number of (innocuous) warnings, and remove a useless test. There are still several signed/unsigned warnings left, which require a bit more study for a proper fix. This file has grown beyond reasonable limits. We really need to split it into separate components (ipv4, ipv6, dummynet, nat, table, userland-kernel communication ...) so we can make mainteinance easier. MFC after: 1 weeks
2009-01-20 18:16:31 +00:00
if ( !(*flags & HAVE_PROTO) && (want & HAVE_PROTO)) {
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
if ( (*flags & HAVE_PROTO4))
printf(" ip4");
else if ( (*flags & HAVE_PROTO6))
printf(" ip6");
else
printf(" ip");
Fix a number of (innocuous) warnings, and remove a useless test. There are still several signed/unsigned warnings left, which require a bit more study for a proper fix. This file has grown beyond reasonable limits. We really need to split it into separate components (ipv4, ipv6, dummynet, nat, table, userland-kernel communication ...) so we can make mainteinance easier. MFC after: 1 weeks
2009-01-20 18:16:31 +00:00
}
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
if ( !(*flags & HAVE_SRCIP) && (want & HAVE_SRCIP))
printf(" from any");
if ( !(*flags & HAVE_DSTIP) && (want & HAVE_DSTIP))
printf(" to any");
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*flags |= want;
}
static void
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
show_ipfw(struct ip_fw *rule, int pcwidth, int bcwidth)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
Align timestamps when -t is used in ipfw and ipfw2. PR: kern/44843 Approved by: re (jhb)
2002-11-26 22:53:14 +00:00
static int twidth = 0;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int l;
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
ipfw_insn *cmd, *tagptr = NULL;
Fix a number of (innocuous) warnings, and remove a useless test. There are still several signed/unsigned warnings left, which require a bit more study for a proper fix. This file has grown beyond reasonable limits. We really need to split it into separate components (ipv4, ipv6, dummynet, nat, table, userland-kernel communication ...) so we can make mainteinance easier. MFC after: 1 weeks
2009-01-20 18:16:31 +00:00
const char *comment = NULL; /* ptr to comment if we have one */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int proto = 0; /* default */
int flags = 0; /* prerequisites */
ipfw_insn_log *logptr = NULL; /* set if we find an O_LOG */
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
ipfw_insn_altq *altqptr = NULL; /* set if we find an O_ALTQ */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int or_block = 0; /* we are in an or block */
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint32_t set_disable;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
bcopy(&rule->next_rule, &set_disable, sizeof(set_disable));
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
if (set_disable & (1 << rule->set)) { /* disabled */
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (!co.show_sets)
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
return;
else
printf("# DISABLED ");
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
printf("%05u ", rule->rulenum);
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
if (pcwidth>0 || bcwidth>0)
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
printf("%*llu %*llu ", pcwidth, align_uint64(&rule->pcnt),
bcwidth, align_uint64(&rule->bcnt));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.do_time == 2)
Add a '-T' flag to print the timestamp as numeric value instead of converting it with ctime(). This is a lot more convenient for postprocessing. Submitted by: "Jacob S. Barrett" <jbarrett@amduat.net>
2003-07-12 08:35:25 +00:00
printf("%10u ", rule->timestamp);
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
else if (co.do_time == 1) {
Align timestamps when -t is used in ipfw and ipfw2. PR: kern/44843 Approved by: re (jhb)
2002-11-26 22:53:14 +00:00
char timestr[30];
time_t t = (time_t)0;
if (twidth == 0) {
strcpy(timestr, ctime(&t));
*strchr(timestr, '\n') = '\0';
twidth = strlen(timestr);
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (rule->timestamp) {
Align timestamps when -t is used in ipfw and ipfw2. PR: kern/44843 Approved by: re (jhb)
2002-11-26 22:53:14 +00:00
t = _long_to_time(rule->timestamp);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
strcpy(timestr, ctime(&t));
*strchr(timestr, '\n') = '\0';
printf("%s ", timestr);
} else {
Align timestamps when -t is used in ipfw and ipfw2. PR: kern/44843 Approved by: re (jhb)
2002-11-26 22:53:14 +00:00
printf("%*s", twidth, " ");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
}
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.show_sets)
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
printf("set %d ", rule->set);
Fix a kernel panic with rules of the type prob 0.5 pipe NN .... due to the generation of an invalid ipfw instruction sequence. No ABI change, but you need to upgrade /sbin/ipfw to generate the correct code. Approved by: re
2002-11-26 19:58:12 +00:00
/*
* print the optional "match probability"
*/
if (rule->cmd_len > 0) {
cmd = rule->cmd ;
if (cmd->opcode == O_PROB) {
ipfw_insn_u32 *p = (ipfw_insn_u32 *)cmd;
double d = 1.0 * p->d[0];
d = (d / 0x7fffffff);
printf("prob %f ", d);
}
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* first print actions
*/
for (l = rule->cmd_len - rule->act_ofs, cmd = ACTION_PTR(rule);
l > 0 ; l -= F_LEN(cmd), cmd += F_LEN(cmd)) {
switch(cmd->opcode) {
case O_CHECK_STATE:
printf("check-state");
print correctly commands of the form ipfw add 100 allow ip from { 1.2.3.4 or 5.6.7.8 } (note that the above example could be better written as ipfw add 100 allow dst-ip 1.2.3.4,5.6.7.8 Submitted by: Riccardo Panicucci
2010-03-15 18:20:51 +00:00
/* avoid printing anything else */
flags = HAVE_PROTO | HAVE_SRCIP |
HAVE_DSTIP | HAVE_IP;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case O_ACCEPT:
printf("allow");
break;
case O_COUNT:
printf("count");
break;
case O_DENY:
printf("deny");
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case O_REJECT:
if (cmd->arg1 == ICMP_REJECT_RST)
printf("reset");
else if (cmd->arg1 == ICMP_UNREACH_HOST)
printf("reject");
else
print_reject_code(cmd->arg1);
break;
* Add dynamic sysctl for net.inet6.ip6.fw. * Correct handling of IPv6 Extension Headers. * Add unreach6 code. * Add logging for IPv6. Submitted by: sysctl handling derived from patch from ume needed for ip6fw Obtained from: is_icmp6_query and send_reject6 derived from similar functions of netinet6,ip6fw Reviewed by: ume, gnn; silence on ipfw@ Test setup provided by: CK Software GmbH MFC after: 6 days
2005-08-13 11:02:34 +00:00
case O_UNREACH6:
if (cmd->arg1 == ICMP6_UNREACH_RST)
printf("reset6");
else
print_unreach6_code(cmd->arg1);
break;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
case O_SKIPTO:
PRINT_UINT_ARG("skipto ", cmd->arg1);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case O_PIPE:
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG("pipe ", cmd->arg1);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_QUEUE:
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG("queue ", cmd->arg1);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_DIVERT:
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG("divert ", cmd->arg1);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_TEE:
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG("tee ", cmd->arg1);
break;
Add a ng_ipfw node, implementing a quick and simple interface between ipfw(4) and netgraph(4) facilities. Reviewed by: andre, brooks, julian
2005-02-05 12:06:33 +00:00
case O_NETGRAPH:
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG("netgraph ", cmd->arg1);
break;
Add a ng_ipfw node, implementing a quick and simple interface between ipfw(4) and netgraph(4) facilities. Reviewed by: andre, brooks, julian
2005-02-05 12:06:33 +00:00
case O_NGTEE:
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG("ngtee ", cmd->arg1);
break;
Add a ng_ipfw node, implementing a quick and simple interface between ipfw(4) and netgraph(4) facilities. Reviewed by: andre, brooks, julian
2005-02-05 12:06:33 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_FORWARD_IP:
{
ipfw_insn_sa *s = (ipfw_insn_sa *)cmd;
Allow ipfw to forward to a destination that is specified by a table. for example: fwd tablearg ip from any to table(1) where table 1 has entries of the form: 1.1.1.0/24 10.2.3.4 208.23.2.0/24 router2 This allows trivial implementation of a secondary routing table implemented in the firewall layer. I expect more work (under discussion with Glebius) to follow this to clean up some of the messy parts of ipfw related to tables. Reviewed by: Glebius MFC after: 1 month
2006-08-17 22:49:50 +00:00
if (s->sa.sin_addr.s_addr == INADDR_ANY) {
printf("fwd tablearg");
} else {
printf("fwd %s", inet_ntoa(s->sa.sin_addr));
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (s->sa.sin_port)
Store the port number in "fwd" rules in host format, same as ipfw1 has always done. Technically, this is the wrong format, but it reduces the diffs in -stable. Someday, when we get rid of ipfw1, I will put the port number in the proper format both in kernel and userland. MFC after: 3 days (with re@ permission)
2002-09-12 00:45:32 +00:00
printf(",%d", s->sa.sin_port);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
break;
case O_LOG: /* O_LOG is printed last */
logptr = (ipfw_insn_log *)cmd;
break;
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
case O_ALTQ: /* O_ALTQ is printed after O_LOG */
altqptr = (ipfw_insn_altq *)cmd;
break;
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
case O_TAG:
tagptr = cmd;
break;
Summer of Code 2005: improve libalias - part 2 of 2 With the second (and last) part of my previous Summer of Code work, we get: -ipfw's in kernel nat -redirect_* and LSNAT support General information about nat syntax and some examples are available in the ipfw (8) man page. The redirect and LSNAT syntax are identical to natd, so please refer to natd (8) man page. To enable in kernel nat in rc.conf, two options were added: o firewall_nat_enable: equivalent to natd_enable o firewall_nat_interface: equivalent to natd_interface Remember to set net.inet.ip.fw.one_pass to 0, if you want the packet to continue being checked by the firewall ruleset after being (de)aliased. NOTA BENE: due to some problems with libalias architecture, in kernel nat won't work with TSO enabled nic, thus you have to disable TSO via ifconfig (ifconfig foo0 -tso). Approved by: glebius (mentor)
2006-12-29 21:59:17 +00:00
case O_NAT:
Add table/tablearg support to ipfw's nat. MFC After: 1 week
2008-02-24 15:37:45 +00:00
PRINT_UINT_ARG("nat ", cmd->arg1);
Summer of Code 2005: improve libalias - part 2 of 2 With the second (and last) part of my previous Summer of Code work, we get: -ipfw's in kernel nat -redirect_* and LSNAT support General information about nat syntax and some examples are available in the ipfw (8) man page. The redirect and LSNAT syntax are identical to natd, so please refer to natd (8) man page. To enable in kernel nat in rc.conf, two options were added: o firewall_nat_enable: equivalent to natd_enable o firewall_nat_interface: equivalent to natd_interface Remember to set net.inet.ip.fw.one_pass to 0, if you want the packet to continue being checked by the firewall ruleset after being (de)aliased. NOTA BENE: due to some problems with libalias architecture, in kernel nat won't work with TSO enabled nic, thus you have to disable TSO via ifconfig (ifconfig foo0 -tso). Approved by: glebius (mentor)
2006-12-29 21:59:17 +00:00
break;
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
case O_SETFIB:
PRINT_UINT_ARG("setfib ", cmd->arg1);
break;
Implement an ipfw action to reassemble ip packets: reass.
2009-04-01 20:23:47 +00:00
case O_REASS:
printf("reass");
break;
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
default:
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
printf("** unrecognized action %d len %d ",
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->opcode, cmd->len);
}
}
if (logptr) {
if (logptr->max_log > 0)
A bunch of minor fixes: * accept "icmptype" as an alias for "icmptypes"; * remove an extra whitespace after "log" rules; * print correctly the "limit" masks; * correct a typo in parsing dummynet arguments (this caused a coredump); * do not allow specifying both "check-state" and "limit", they are (and have always been) mutually exclusive; * remove an extra print of the rule before installing it; * make stdout buffered -- otherwise, if you log its output with syslog, you will see one entry for each printf(). Rather unpleasant.
2002-07-13 15:57:23 +00:00
printf(" log logamount %d", logptr->max_log);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
else
A bunch of minor fixes: * accept "icmptype" as an alias for "icmptypes"; * remove an extra whitespace after "log" rules; * print correctly the "limit" masks; * correct a typo in parsing dummynet arguments (this caused a coredump); * do not allow specifying both "check-state" and "limit", they are (and have always been) mutually exclusive; * remove an extra print of the rule before installing it; * make stdout buffered -- otherwise, if you log its output with syslog, you will see one entry for each printf(). Rather unpleasant.
2002-07-13 15:57:23 +00:00
printf(" log");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#ifndef NO_ALTQ
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
if (altqptr) {
put the altq-related functions into a separate file. Minor cleanup of the includes used by the various source files, including annotations of why certain headers are used.
2009-02-01 16:00:49 +00:00
print_altq_cmd(altqptr);
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
}
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#endif
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
if (tagptr) {
if (tagptr->len & F_NOT)
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG(" untag ", tagptr->arg1);
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
else
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG(" tag ", tagptr->arg1);
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
}
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
* then print the body.
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
for (l = rule->act_ofs, cmd = rule->cmd ;
l > 0 ; l -= F_LEN(cmd) , cmd += F_LEN(cmd)) {
if ((cmd->len & F_OR) || (cmd->len & F_NOT))
continue;
if (cmd->opcode == O_IP4) {
flags |= HAVE_PROTO4;
break;
} else if (cmd->opcode == O_IP6) {
flags |= HAVE_PROTO6;
break;
}
}
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
if (rule->_pad & 1) { /* empty rules before options */
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (!co.do_compact) {
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
show_prerequisites(&flags, HAVE_PROTO, 0);
printf(" from any to any");
}
print correctly commands of the form ipfw add 100 allow ip from { 1.2.3.4 or 5.6.7.8 } (note that the above example could be better written as ipfw add 100 allow dst-ip 1.2.3.4,5.6.7.8 Submitted by: Riccardo Panicucci
2010-03-15 18:20:51 +00:00
flags |= HAVE_IP | HAVE_OPTIONS | HAVE_PROTO |
HAVE_SRCIP | HAVE_DSTIP;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
}
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.comment_only)
Add a -b flag to /sbin/ipfw to print only action and comment for each rule, thus omitting the entire body. This makes the output a lot more readable for complex rulesets (provided, of course, you have annotated your ruleset appropriately!) MFC after: 3 days
2003-12-12 16:14:28 +00:00
comment = "...";
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
for (l = rule->act_ofs, cmd = rule->cmd ;
l > 0 ; l -= F_LEN(cmd) , cmd += F_LEN(cmd)) {
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
/* useful alias */
ipfw_insn_u32 *cmd32 = (ipfw_insn_u32 *)cmd;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.comment_only) {
Add a -b flag to /sbin/ipfw to print only action and comment for each rule, thus omitting the entire body. This makes the output a lot more readable for complex rulesets (provided, of course, you have annotated your ruleset appropriately!) MFC after: 3 days
2003-12-12 16:14:28 +00:00
if (cmd->opcode != O_NOP)
continue;
printf(" // %s\n", (char *)(cmd + 1));
return;
}
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
show_prerequisites(&flags, 0, cmd->opcode);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
switch(cmd->opcode) {
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
case O_PROB:
Fix a kernel panic with rules of the type prob 0.5 pipe NN .... due to the generation of an invalid ipfw instruction sequence. No ABI change, but you need to upgrade /sbin/ipfw to generate the correct code. Approved by: re
2002-11-26 19:58:12 +00:00
break; /* done already */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_PROBE_STATE:
break; /* no need to print anything here */
case O_IP_SRC:
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
case O_IP_SRC_LOOKUP:
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_SRC_MASK:
case O_IP_SRC_ME:
case O_IP_SRC_SET:
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
show_prerequisites(&flags, HAVE_PROTO, 0);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (!(flags & HAVE_SRCIP))
printf(" from");
if ((cmd->len & F_OR) && !or_block)
printf(" {");
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
print_ip((ipfw_insn_ip *)cmd,
(flags & HAVE_OPTIONS) ? " src-ip" : "");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
flags |= HAVE_SRCIP;
break;
case O_IP_DST:
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
case O_IP_DST_LOOKUP:
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_DST_MASK:
case O_IP_DST_ME:
case O_IP_DST_SET:
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
show_prerequisites(&flags, HAVE_PROTO|HAVE_SRCIP, 0);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (!(flags & HAVE_DSTIP))
printf(" to");
if ((cmd->len & F_OR) && !or_block)
printf(" {");
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
print_ip((ipfw_insn_ip *)cmd,
(flags & HAVE_OPTIONS) ? " dst-ip" : "");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
flags |= HAVE_DSTIP;
break;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
case O_IP6_SRC:
case O_IP6_SRC_MASK:
case O_IP6_SRC_ME:
add_proto() now fills proto for us so stop to 'guess' the protocol from the command and rather trust the value add_proto filled in. While here, fix an oversight in the pretty printing of ip6/4 options.
2005-06-07 14:11:17 +00:00
show_prerequisites(&flags, HAVE_PROTO, 0);
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
if (!(flags & HAVE_SRCIP))
printf(" from");
if ((cmd->len & F_OR) && !or_block)
printf(" {");
print_ip6((ipfw_insn_ip6 *)cmd,
(flags & HAVE_OPTIONS) ? " src-ip6" : "");
flags |= HAVE_SRCIP | HAVE_PROTO;
break;
case O_IP6_DST:
case O_IP6_DST_MASK:
case O_IP6_DST_ME:
show_prerequisites(&flags, HAVE_PROTO|HAVE_SRCIP, 0);
if (!(flags & HAVE_DSTIP))
printf(" to");
if ((cmd->len & F_OR) && !or_block)
printf(" {");
print_ip6((ipfw_insn_ip6 *)cmd,
(flags & HAVE_OPTIONS) ? " dst-ip6" : "");
flags |= HAVE_DSTIP;
break;
case O_FLOW6ID:
print_flow6id( (ipfw_insn_u32 *) cmd );
flags |= HAVE_OPTIONS;
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_DSTPORT:
print correctly commands of the form ipfw add 100 allow ip from { 1.2.3.4 or 5.6.7.8 } (note that the above example could be better written as ipfw add 100 allow dst-ip 1.2.3.4,5.6.7.8 Submitted by: Riccardo Panicucci
2010-03-15 18:20:51 +00:00
show_prerequisites(&flags,
HAVE_PROTO | HAVE_SRCIP |
HAVE_DSTIP | HAVE_IP, 0);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_SRCPORT:
print correctly commands of the form ipfw add 100 allow ip from { 1.2.3.4 or 5.6.7.8 } (note that the above example could be better written as ipfw add 100 allow dst-ip 1.2.3.4,5.6.7.8 Submitted by: Riccardo Panicucci
2010-03-15 18:20:51 +00:00
show_prerequisites(&flags,
HAVE_PROTO | HAVE_SRCIP, 0);
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
if ((cmd->len & F_OR) && !or_block)
printf(" {");
o Cosmetic: fix the issue when "ipfw(8) show" produces "not" twice: $ ipfw -n add 1 allow layer2 not mac-type ip 00001 allow ip from any to any layer2 not not mac-type 0x0800 PR: bin/115372 Submitted by: Andrey V. Elsukov Approved by: re (hrs) MFC after: 3 weeks
2007-09-23 16:29:22 +00:00
if (cmd->len & F_NOT)
printf(" not");
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
print_newports((ipfw_insn_u16 *)cmd, proto,
(flags & HAVE_OPTIONS) ? cmd->opcode : 0);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case O_PROTO: {
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
struct protoent *pe = NULL;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if ((cmd->len & F_OR) && !or_block)
printf(" {");
if (cmd->len & F_NOT)
printf(" not");
proto = cmd->arg1;
Fix a the previous commit. I wanted to remove the if and always run the body not remove both. Reported by: ceri Pointy hat: brooks
2005-04-26 20:22:31 +00:00
pe = getprotobynumber(cmd->arg1);
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
if ((flags & (HAVE_PROTO4 | HAVE_PROTO6)) &&
!(flags & HAVE_PROTO))
show_prerequisites(&flags,
print correctly commands of the form ipfw add 100 allow ip from { 1.2.3.4 or 5.6.7.8 } (note that the above example could be better written as ipfw add 100 allow dst-ip 1.2.3.4,5.6.7.8 Submitted by: Riccardo Panicucci
2010-03-15 18:20:51 +00:00
HAVE_PROTO | HAVE_IP | HAVE_SRCIP |
HAVE_DSTIP | HAVE_OPTIONS, 0);
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
if (flags & HAVE_OPTIONS)
printf(" proto");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (pe)
printf(" %s", pe->p_name);
else
printf(" %u", cmd->arg1);
}
flags |= HAVE_PROTO;
break;
Kill EOL whitespaces, style(9) fix.
2002-11-06 15:09:34 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
default: /*options ... */
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
if (!(cmd->len & (F_OR|F_NOT)))
if (((cmd->opcode == O_IP6) &&
(flags & HAVE_PROTO6)) ||
((cmd->opcode == O_IP4) &&
(flags & HAVE_PROTO4)))
break;
print correctly commands of the form ipfw add 100 allow ip from { 1.2.3.4 or 5.6.7.8 } (note that the above example could be better written as ipfw add 100 allow dst-ip 1.2.3.4,5.6.7.8 Submitted by: Riccardo Panicucci
2010-03-15 18:20:51 +00:00
show_prerequisites(&flags, HAVE_PROTO | HAVE_SRCIP |
HAVE_DSTIP | HAVE_IP | HAVE_OPTIONS, 0);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if ((cmd->len & F_OR) && !or_block)
printf(" {");
if (cmd->len & F_NOT && cmd->opcode != O_IN)
printf(" not");
switch(cmd->opcode) {
o Make ipfw(8) show rules with mac/mac-type options correctly. Before: $ ipfw -n add 100 count icmp from any to any mac-type 0x01 00100 count icmp 0x0001 $ ipfw -n add 100 count icmp from any to any mac any any 00100 count icmp MAC any any any After: $ ipfw -n add 100 count icmp from any to any mac-type 0x01 00100 count icmp from any to any mac-type 0x0001 $ ipfw -n add 100 count icmp from any to any mac any any 00100 count icmp from any to any MAC any any PR: bin/112244 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-04-30 17:39:30 +00:00
case O_MACADDR2: {
ipfw_insn_mac *m = (ipfw_insn_mac *)cmd;
printf(" MAC");
print_mac(m->addr, m->mask);
print_mac(m->addr + 6, m->mask + 6);
}
break;
case O_MAC_TYPE:
print_newports((ipfw_insn_u16 *)cmd,
IPPROTO_ETHERTYPE, cmd->opcode);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_FRAG:
printf(" frag");
break;
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
case O_FIB:
printf(" fib %u", cmd->arg1 );
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IN:
printf(cmd->len & F_NOT ? " out" : " in");
break;
Add support to IPFW for classification based on "diverted" status (that is, input via a divert socket).
2004-10-03 00:26:35 +00:00
case O_DIVERTED:
switch (cmd->arg1) {
case 3:
printf(" diverted");
break;
case 1:
printf(" diverted-loopback");
break;
case 2:
printf(" diverted-output");
break;
default:
printf(" diverted-?<%u>", cmd->arg1);
break;
}
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_LAYER2:
printf(" layer2");
break;
case O_XMIT:
case O_RECV:
Don't print extra " via ", if we have already printed one. While here, slightly style brackets. PR: misc/75297 MFC after: 1 week
2005-01-18 15:13:20 +00:00
case O_VIA:
{
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
char const *s;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
ipfw_insn_if *cmdif = (ipfw_insn_if *)cmd;
if (cmd->opcode == O_XMIT)
s = "xmit";
else if (cmd->opcode == O_RECV)
s = "recv";
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
else /* if (cmd->opcode == O_VIA) */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
s = "via";
if (cmdif->name[0] == '\0')
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
printf(" %s %s", s,
inet_ntoa(cmdif->p.ip));
Don't print extra " via ", if we have already printed one. While here, slightly style brackets. PR: misc/75297 MFC after: 1 week
2005-01-18 15:13:20 +00:00
else
printf(" %s %s", s, cmdif->name);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Don't print extra " via ", if we have already printed one. While here, slightly style brackets. PR: misc/75297 MFC after: 1 week
2005-01-18 15:13:20 +00:00
break;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IPID:
Add support for multiple values and ranges for the "iplen", "ipttl", "ipid" options. This feature has been requested by several users. On passing, fix some minor bugs in the parser. This change is fully backward compatible so if you have an old /sbin/ipfw and a new kernel you are not in trouble (but you need to update /sbin/ipfw if you want to use the new features). Document the changes in the manpage. Now you can write things like ipfw add skipto 1000 iplen 0-500 which some people were asking to give preferential treatment to short packets. The 'MFC after' is just set as a reminder, because I still need to merge the Alpha/Sparc64 fixes for ipfw2 (which unfortunately change the size of certain kernel structures; not that it matters a lot since ipfw2 is entirely optional and not the default...) PR: bin/48015 MFC after: 1 week
2003-06-22 17:33:19 +00:00
if (F_LEN(cmd) == 1)
printf(" ipid %u", cmd->arg1 );
else
print_newports((ipfw_insn_u16 *)cmd, 0,
O_IPID);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case O_IPTTL:
Add support for multiple values and ranges for the "iplen", "ipttl", "ipid" options. This feature has been requested by several users. On passing, fix some minor bugs in the parser. This change is fully backward compatible so if you have an old /sbin/ipfw and a new kernel you are not in trouble (but you need to update /sbin/ipfw if you want to use the new features). Document the changes in the manpage. Now you can write things like ipfw add skipto 1000 iplen 0-500 which some people were asking to give preferential treatment to short packets. The 'MFC after' is just set as a reminder, because I still need to merge the Alpha/Sparc64 fixes for ipfw2 (which unfortunately change the size of certain kernel structures; not that it matters a lot since ipfw2 is entirely optional and not the default...) PR: bin/48015 MFC after: 1 week
2003-06-22 17:33:19 +00:00
if (F_LEN(cmd) == 1)
printf(" ipttl %u", cmd->arg1 );
else
print_newports((ipfw_insn_u16 *)cmd, 0,
O_IPTTL);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case O_IPVER:
printf(" ipver %u", cmd->arg1 );
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case O_IPPRECEDENCE:
printf(" ipprecedence %u", (cmd->arg1) >> 5 );
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IPLEN:
Add support for multiple values and ranges for the "iplen", "ipttl", "ipid" options. This feature has been requested by several users. On passing, fix some minor bugs in the parser. This change is fully backward compatible so if you have an old /sbin/ipfw and a new kernel you are not in trouble (but you need to update /sbin/ipfw if you want to use the new features). Document the changes in the manpage. Now you can write things like ipfw add skipto 1000 iplen 0-500 which some people were asking to give preferential treatment to short packets. The 'MFC after' is just set as a reminder, because I still need to merge the Alpha/Sparc64 fixes for ipfw2 (which unfortunately change the size of certain kernel structures; not that it matters a lot since ipfw2 is entirely optional and not the default...) PR: bin/48015 MFC after: 1 week
2003-06-22 17:33:19 +00:00
if (F_LEN(cmd) == 1)
printf(" iplen %u", cmd->arg1 );
else
print_newports((ipfw_insn_u16 *)cmd, 0,
O_IPLEN);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
case O_IPOPT:
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
print_flags("ipoptions", cmd, f_ipopts);
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case O_IPTOS:
print_flags("iptos", cmd, f_iptos);
break;
case O_ICMPTYPE:
print_icmptypes((ipfw_insn_u32 *)cmd);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_ESTAB:
printf(" established");
break;
Add support to IPFW for matching by TCP data length.
2004-10-03 00:47:15 +00:00
case O_TCPDATALEN:
if (F_LEN(cmd) == 1)
printf(" tcpdatalen %u", cmd->arg1 );
else
print_newports((ipfw_insn_u16 *)cmd, 0,
O_TCPDATALEN);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_TCPFLAGS:
print_flags("tcpflags", cmd, f_tcpflags);
break;
case O_TCPOPTS:
print_flags("tcpoptions", cmd, f_tcpopts);
break;
case O_TCPWIN:
printf(" tcpwin %d", ntohs(cmd->arg1));
break;
case O_TCPACK:
printf(" tcpack %d", ntohl(cmd32->d[0]));
break;
case O_TCPSEQ:
printf(" tcpseq %d", ntohl(cmd32->d[0]));
break;
case O_UID:
{
struct passwd *pwd = getpwuid(cmd32->d[0]);
if (pwd)
printf(" uid %s", pwd->pw_name);
else
printf(" uid %u", cmd32->d[0]);
}
break;
case O_GID:
{
struct group *grp = getgrgid(cmd32->d[0]);
if (grp)
printf(" gid %s", grp->gr_name);
else
printf(" gid %u", cmd32->d[0]);
}
break;
Add the ability to associate ipfw rules with a specific prison ID. Since the only thing truly unique about a prison is it's ID, I figured this would be the most granular way of handling this. This commit makes the following changes: - Adds tokenizing and parsing for the ``jail'' command line option to the ipfw(8) userspace utility. - Append the ipfw opcode list with O_JAIL. - While Iam here, add a comment informing others that if they want to add additional opcodes, they should append them to the end of the list to avoid ABI breakage. - Add ``fw_prid'' to the ipfw ucred cache structure. - When initializing ucred cache, if the process is jailed, set fw_prid to the prison ID, otherwise set it to -1. - Update man page to reflect these changes. This change was a strong motivator behind the ucred caching mechanism in ipfw. A sample usage of this new functionality could be: ipfw add count ip from any to any jail 2 It should be noted that because ucred based constraints are only implemented for TCP and UDP packets, the same applies for jail associations. Conceptual head nod by: pjd Reviewed by: rwatson Approved by: bmilekic (mentor)
2004-08-12 22:06:55 +00:00
case O_JAIL:
printf(" jail %d", cmd32->d[0]);
break;
Add a 'verrevpath' option that verifies the interface that a packet comes in on is the same interface that we would route out of to get to the packet's source address. Essentially automates an anti-spoofing check using the information in the routing table. Experimental. The usage and rule format for the feature may still be subject to change.
2003-03-15 01:13:00 +00:00
case O_VERREVPATH:
printf(" verrevpath");
break;
remove extra whitespace and blank lines
2003-06-27 17:18:14 +00:00
Add the option versrcreach to verify that a valid route to the source address of a packet exists in the routing table. The default route is ignored because it would match everything and render the check pointless. This option is very useful for routers with a complete view of the Internet (BGP) in the routing table to reject packets with spoofed or unrouteable source addresses. Example: ipfw add 1000 deny ip from any to any not versrcreach also known in Cisco-speak as: ip verify unicast source reachable-via any Reviewed by: luigi
2004-04-23 14:28:38 +00:00
case O_VERSRCREACH:
printf(" versrcreach");
break;
New ipfw option "antispoof": For incoming packets, the packet's source address is checked if it belongs to a directly connected network. If the network is directly connected, then the interface the packet came on in is compared to the interface the network is connected to. When incoming interface and directly connected interface are not the same, the packet does not match. Usage example: ipfw add deny ip from any to any not antispoof in Manpage education by: ru
2004-08-09 16:12:10 +00:00
case O_ANTISPOOF:
printf(" antispoof");
break;
Implement the 'ipsec' option to match packets coming out of an ipsec tunnel. Should work with both regular and fast ipsec (mutually exclusive). See manpage for more details. Submitted by: Ari Suutari (ari.suutari@syncrontech.com) Revised by: sam MFC after: 1 week
2003-07-04 21:42:32 +00:00
case O_IPSEC:
printf(" ipsec");
break;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
case O_NOP:
Make sure that comments are printed at the end of a rule. Reported by: Patrick Tracanelli <eksffa@freebsdbrasil.com.br>
2003-07-15 10:23:43 +00:00
comment = (char *)(cmd + 1);
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_KEEP_STATE:
printf(" keep-state");
break;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
case O_LIMIT: {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
struct _s_x *p = limit_masks;
ipfw_insn_limit *c = (ipfw_insn_limit *)cmd;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint8_t x = c->limit_mask;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
char const *comma = " ";
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
printf(" limit");
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
for (; p->x != 0 ; p++)
A bunch of minor fixes: * accept "icmptype" as an alias for "icmptypes"; * remove an extra whitespace after "log" rules; * print correctly the "limit" masks; * correct a typo in parsing dummynet arguments (this caused a coredump); * do not allow specifying both "check-state" and "limit", they are (and have always been) mutually exclusive; * remove an extra print of the rule before installing it; * make stdout buffered -- otherwise, if you log its output with syslog, you will see one entry for each printf(). Rather unpleasant.
2002-07-13 15:57:23 +00:00
if ((x & p->x) == p->x) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
x &= ~p->x;
printf("%s%s", comma, p->s);
comma = ",";
}
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG(" ", c->conn_limit);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
case O_IP6:
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
printf(" ip6");
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
break;
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
case O_IP4:
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
printf(" ip4");
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
break;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
case O_ICMP6TYPE:
print_icmp6types((ipfw_insn_u32 *)cmd);
break;
case O_EXT_HDR:
print_ext6hdr( (ipfw_insn *) cmd );
break;
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
case O_TAGGED:
if (F_LEN(cmd) == 1)
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
PRINT_UINT_ARG(" tagged ", cmd->arg1);
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
else
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
print_newports((ipfw_insn_u16 *)cmd, 0,
O_TAGGED);
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
default:
printf(" [opcode %d len %d]",
cmd->opcode, cmd->len);
}
}
if (cmd->len & F_OR) {
printf(" or");
or_block = 1;
} else if (or_block) {
printf(" }");
or_block = 0;
}
}
print correctly commands of the form ipfw add 100 allow ip from { 1.2.3.4 or 5.6.7.8 } (note that the above example could be better written as ipfw add 100 allow dst-ip 1.2.3.4,5.6.7.8 Submitted by: Riccardo Panicucci
2010-03-15 18:20:51 +00:00
show_prerequisites(&flags, HAVE_PROTO | HAVE_SRCIP | HAVE_DSTIP
| HAVE_IP, 0);
Make sure that comments are printed at the end of a rule. Reported by: Patrick Tracanelli <eksffa@freebsdbrasil.com.br>
2003-07-15 10:23:43 +00:00
if (comment)
printf(" // %s", comment);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
printf("\n");
}
static void
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
show_dyn_ipfw(ipfw_dyn_rule *d, int pcwidth, int bcwidth)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
struct protoent *pe;
struct in_addr a;
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
uint16_t rulenum;
Print dynamic rules for IPv6 as well. PR: bin/98349 Submitted by: Mark Andrews MFC after: 2 weeks
2006-06-02 05:17:17 +00:00
char buf[INET6_ADDRSTRLEN];
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (!co.do_expired) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (!d->expire && !(d->dyn_type == O_LIMIT_PARENT))
return;
}
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
bcopy(&d->rule, &rulenum, sizeof(rulenum));
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
printf("%05d", rulenum);
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
if (pcwidth>0 || bcwidth>0)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
printf(" %*llu %*llu (%ds)", pcwidth,
align_uint64(&d->pcnt), bcwidth,
align_uint64(&d->bcnt), d->expire);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
switch (d->dyn_type) {
case O_LIMIT_PARENT:
printf(" PARENT %d", d->count);
break;
case O_LIMIT:
printf(" LIMIT");
break;
case O_KEEP_STATE: /* bidir, no mask */
Kill EOL whitespaces, style(9) fix.
2002-11-06 15:09:34 +00:00
printf(" STATE");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
}
if ((pe = getprotobynumber(d->id.proto)) != NULL)
printf(" %s", pe->p_name);
else
printf(" proto %u", d->id.proto);
Print dynamic rules for IPv6 as well. PR: bin/98349 Submitted by: Mark Andrews MFC after: 2 weeks
2006-06-02 05:17:17 +00:00
if (d->id.addr_type == 4) {
a.s_addr = htonl(d->id.src_ip);
printf(" %s %d", inet_ntoa(a), d->id.src_port);
a.s_addr = htonl(d->id.dst_ip);
printf(" <-> %s %d", inet_ntoa(a), d->id.dst_port);
} else if (d->id.addr_type == 6) {
printf(" %s %d", inet_ntop(AF_INET6, &d->id.src_ip6, buf,
sizeof(buf)), d->id.src_port);
printf(" <-> %s %d", inet_ntop(AF_INET6, &d->id.dst_ip6, buf,
sizeof(buf)), d->id.dst_port);
} else
printf(" UNKNOWN <-> UNKNOWN\n");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
printf("\n");
}
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
/*
* This one handles all set-related commands
* ipfw set { show | enable | disable }
* ipfw set swap X Y
* ipfw set move X to Y
* ipfw set move rule X to Y
*/
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
void
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
ipfw_sets_handler(char *av[])
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
{
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint32_t set_disable, masks[2];
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
int i, nbytes;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint16_t rulenum;
uint8_t cmd, new_set;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
av++;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] == NULL)
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
errx(EX_USAGE, "set needs command");
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(*av, "show") == 0) {
fix handling of sets
2010-03-04 16:55:32 +00:00
void *data = NULL;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
char const *msg;
fix handling of sets
2010-03-04 16:55:32 +00:00
int nalloc;
nalloc = nbytes = sizeof(struct ip_fw);
while (nbytes >= nalloc) {
if (data)
free(data);
nalloc = nalloc * 2 + 200;
nbytes = nalloc;
wrap all malloc/calloc/realloc calls so they exit on failure without having to check in each place. Remove an wrong strdup from previous commit.
2009-01-26 14:26:35 +00:00
data = safe_calloc(1, nbytes);
Apply a bandaid to get this working on sparc64 again; the introduction of do_cmd() broke things, because this function assumes that a socklen_t is large enough to hold a pointer. A real solution to this problem would be a rewrite of do_cmd() to treat the optlen parameter consistently and not use it to carry a pointer or integer dependent on the context.
2003-09-04 15:57:37 +00:00
if (do_cmd(IP_FW_GET, data, (uintptr_t)&nbytes) < 0)
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
err(EX_OSERR, "getsockopt(IP_FW_GET)");
fix handling of sets
2010-03-04 16:55:32 +00:00
}
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
bcopy(&((struct ip_fw *)data)->next_rule,
&set_disable, sizeof(set_disable));
Userland side of: Allow set 31 to be used for rules other than 65535. Set 31 is still special because rules belonging to it are not deleted by the "ipfw flush" command, but must be deleted explicitly with "ipfw delete set 31" or by individual rule numbers. This implement a flexible form of "persistent rules" which you might want to have available even after an "ipfw flush". Note that this change does not violate POLA, because you could not use set 31 in a ruleset before this change. Suggested by: Paul Richards
2003-07-15 23:08:44 +00:00
for (i = 0, msg = "disable" ; i < RESVD_SET; i++)
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
if ((set_disable & (1<<i))) {
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
printf("%s %d", msg, i);
msg = "";
}
msg = (set_disable) ? " enable" : "enable";
Userland side of: Allow set 31 to be used for rules other than 65535. Set 31 is still special because rules belonging to it are not deleted by the "ipfw flush" command, but must be deleted explicitly with "ipfw delete set 31" or by individual rule numbers. This implement a flexible form of "persistent rules" which you might want to have available even after an "ipfw flush". Note that this change does not violate POLA, because you could not use set 31 in a ruleset before this change. Suggested by: Paul Richards
2003-07-15 23:08:44 +00:00
for (i = 0; i < RESVD_SET; i++)
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
if (!(set_disable & (1<<i))) {
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
printf("%s %d", msg, i);
msg = "";
}
printf("\n");
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "swap") == 0) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
if ( av[0] == NULL || av[1] == NULL )
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
errx(EX_USAGE, "set swap needs 2 set numbers\n");
rulenum = atoi(av[0]);
new_set = atoi(av[1]);
Userland side of: Allow set 31 to be used for rules other than 65535. Set 31 is still special because rules belonging to it are not deleted by the "ipfw flush" command, but must be deleted explicitly with "ipfw delete set 31" or by individual rule numbers. This implement a flexible form of "persistent rules" which you might want to have available even after an "ipfw flush". Note that this change does not violate POLA, because you could not use set 31 in a ruleset before this change. Suggested by: Paul Richards
2003-07-15 23:08:44 +00:00
if (!isdigit(*(av[0])) || rulenum > RESVD_SET)
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
errx(EX_DATAERR, "invalid set number %s\n", av[0]);
Userland side of: Allow set 31 to be used for rules other than 65535. Set 31 is still special because rules belonging to it are not deleted by the "ipfw flush" command, but must be deleted explicitly with "ipfw delete set 31" or by individual rule numbers. This implement a flexible form of "persistent rules" which you might want to have available even after an "ipfw flush". Note that this change does not violate POLA, because you could not use set 31 in a ruleset before this change. Suggested by: Paul Richards
2003-07-15 23:08:44 +00:00
if (!isdigit(*(av[1])) || new_set > RESVD_SET)
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
errx(EX_DATAERR, "invalid set number %s\n", av[1]);
masks[0] = (4 << 24) | (new_set << 16) | (rulenum);
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
i = do_cmd(IP_FW_DEL, masks, sizeof(uint32_t));
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "move") == 0) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
fix handling of sets
2010-03-04 16:55:32 +00:00
if (av[0] && _substrcmp(*av, "rule") == 0) {
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
cmd = 2;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
} else
cmd = 3;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] == NULL || av[1] == NULL || av[2] == NULL ||
av[3] != NULL || _substrcmp(av[1], "to") != 0)
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
errx(EX_USAGE, "syntax: set move [rule] X to Y\n");
rulenum = atoi(av[0]);
new_set = atoi(av[2]);
Userland side of: Allow set 31 to be used for rules other than 65535. Set 31 is still special because rules belonging to it are not deleted by the "ipfw flush" command, but must be deleted explicitly with "ipfw delete set 31" or by individual rule numbers. This implement a flexible form of "persistent rules" which you might want to have available even after an "ipfw flush". Note that this change does not violate POLA, because you could not use set 31 in a ruleset before this change. Suggested by: Paul Richards
2003-07-15 23:08:44 +00:00
if (!isdigit(*(av[0])) || (cmd == 3 && rulenum > RESVD_SET) ||
Use IPFW_DEFAULT_RULE instead of hardcoded value since now it is available. MFC after: 5 days.
2008-09-06 17:23:37 +00:00
(cmd == 2 && rulenum == IPFW_DEFAULT_RULE) )
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
errx(EX_DATAERR, "invalid source number %s\n", av[0]);
Userland side of: Allow set 31 to be used for rules other than 65535. Set 31 is still special because rules belonging to it are not deleted by the "ipfw flush" command, but must be deleted explicitly with "ipfw delete set 31" or by individual rule numbers. This implement a flexible form of "persistent rules" which you might want to have available even after an "ipfw flush". Note that this change does not violate POLA, because you could not use set 31 in a ruleset before this change. Suggested by: Paul Richards
2003-07-15 23:08:44 +00:00
if (!isdigit(*(av[2])) || new_set > RESVD_SET)
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
errx(EX_DATAERR, "invalid dest. set %s\n", av[1]);
masks[0] = (cmd << 24) | (new_set << 16) | (rulenum);
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
i = do_cmd(IP_FW_DEL, masks, sizeof(uint32_t));
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "disable") == 0 ||
_substrcmp(*av, "enable") == 0 ) {
int which = _substrcmp(*av, "enable") == 0 ? 1 : 0;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
masks[0] = masks[1] = 0;
fix handling of sets
2010-03-04 16:55:32 +00:00
while (av[0]) {
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (isdigit(**av)) {
i = atoi(*av);
Userland side of: Allow set 31 to be used for rules other than 65535. Set 31 is still special because rules belonging to it are not deleted by the "ipfw flush" command, but must be deleted explicitly with "ipfw delete set 31" or by individual rule numbers. This implement a flexible form of "persistent rules" which you might want to have available even after an "ipfw flush". Note that this change does not violate POLA, because you could not use set 31 in a ruleset before this change. Suggested by: Paul Richards
2003-07-15 23:08:44 +00:00
if (i < 0 || i > RESVD_SET)
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
errx(EX_DATAERR,
"invalid set number %d\n", i);
masks[which] |= (1<<i);
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "disable") == 0)
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
which = 0;
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
else if (_substrcmp(*av, "enable") == 0)
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
which = 1;
else
errx(EX_DATAERR,
"invalid set command %s\n", *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
}
if ( (masks[0] & masks[1]) != 0 )
errx(EX_DATAERR,
"cannot enable and disable the same set\n");
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
i = do_cmd(IP_FW_DEL, masks, sizeof(masks));
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (i)
warn("set enable/disable: setsockopt(IP_FW_DEL)");
} else
errx(EX_USAGE, "invalid set command %s\n", *av);
}
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
void
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
ipfw_sysctl_handler(char *av[], int which)
It turns out that we do not need to add a new ioctl to unbreak a default-to-deny firewall. Simply turning off IPFW via a preexisting sysctl does the job. To make it more apparent (since nobody picked up on this in a week's worth of flames), the boolean sysctl's have been integrated into the /sbin/ipfw command set in an obvious and straightforward manner. For example, you can now do 'ipfw disable firewall' or 'ipfw enable firewall'. This is far easier to remember then the net.inet.ip.fw.enable sysctl. Reviewed by: imp MFC after: 3 days
2003-01-12 03:31:10 +00:00
{
av++;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] == NULL) {
It turns out that we do not need to add a new ioctl to unbreak a default-to-deny firewall. Simply turning off IPFW via a preexisting sysctl does the job. To make it more apparent (since nobody picked up on this in a week's worth of flames), the boolean sysctl's have been integrated into the /sbin/ipfw command set in an obvious and straightforward manner. For example, you can now do 'ipfw disable firewall' or 'ipfw enable firewall'. This is far easier to remember then the net.inet.ip.fw.enable sysctl. Reviewed by: imp MFC after: 3 days
2003-01-12 03:31:10 +00:00
warnx("missing keyword to enable/disable\n");
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "firewall") == 0) {
Split some long lines to fit 80 columns (the code in RELENG_4 was already correct).
2003-06-23 22:32:14 +00:00
sysctlbyname("net.inet.ip.fw.enable", NULL, 0,
&which, sizeof(which));
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "one_pass") == 0) {
Split some long lines to fit 80 columns (the code in RELENG_4 was already correct).
2003-06-23 22:32:14 +00:00
sysctlbyname("net.inet.ip.fw.one_pass", NULL, 0,
&which, sizeof(which));
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "debug") == 0) {
Split some long lines to fit 80 columns (the code in RELENG_4 was already correct).
2003-06-23 22:32:14 +00:00
sysctlbyname("net.inet.ip.fw.debug", NULL, 0,
&which, sizeof(which));
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "verbose") == 0) {
Split some long lines to fit 80 columns (the code in RELENG_4 was already correct).
2003-06-23 22:32:14 +00:00
sysctlbyname("net.inet.ip.fw.verbose", NULL, 0,
&which, sizeof(which));
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "dyn_keepalive") == 0) {
Split some long lines to fit 80 columns (the code in RELENG_4 was already correct).
2003-06-23 22:32:14 +00:00
sysctlbyname("net.inet.ip.fw.dyn_keepalive", NULL, 0,
&which, sizeof(which));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#ifndef NO_ALTQ
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "altq") == 0) {
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
altq_set_enabled(which);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#endif
It turns out that we do not need to add a new ioctl to unbreak a default-to-deny firewall. Simply turning off IPFW via a preexisting sysctl does the job. To make it more apparent (since nobody picked up on this in a week's worth of flames), the boolean sysctl's have been integrated into the /sbin/ipfw command set in an obvious and straightforward manner. For example, you can now do 'ipfw disable firewall' or 'ipfw enable firewall'. This is far easier to remember then the net.inet.ip.fw.enable sysctl. Reviewed by: imp MFC after: 3 days
2003-01-12 03:31:10 +00:00
} else {
warnx("unrecognize enable/disable keyword: %s\n", *av);
}
}
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
void
ipfw_list(int ac, char *av[], int show_counters)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
struct ip_fw *r;
ipfw_dyn_rule *dynrules, *d;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
#define NEXT(r) ((struct ip_fw *)((char *)r + RULESIZE(r)))
char *lim;
void *data = NULL;
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
int bcwidth, n, nbytes, nstat, ndyn, pcwidth, width;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int exitval = EX_OK;
int lac;
char **lav;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
u_long rnum, last;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
char *endptr;
int seen = 0;
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
uint8_t set;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
const int ocmd = co.do_pipe ? IP_DUMMYNET_GET : IP_FW_GET;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int nalloc = 1024; /* start somewhere... */
o Initialize a local variable and make gcc happy. PR: bin/71485 Submitted by: Jukka A. Ukkonen
2004-09-10 13:42:24 +00:00
last = 0;
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.test_only) {
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
fprintf(stderr, "Testing only, list disabled\n");
return;
}
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (co.do_pipe) {
dummynet_list(ac, av, show_counters);
return;
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
ac--;
av++;
/* get rules or pipes from kernel, resizing array as necessary */
nbytes = nalloc;
while (nbytes >= nalloc) {
nalloc = nalloc * 2 + 200;
nbytes = nalloc;
wrap all malloc/calloc/realloc calls so they exit on failure without having to check in each place. Remove an wrong strdup from previous commit.
2009-01-26 14:26:35 +00:00
data = safe_realloc(data, nbytes);
Apply a bandaid to get this working on sparc64 again; the introduction of do_cmd() broke things, because this function assumes that a socklen_t is large enough to hold a pointer. A real solution to this problem would be a rewrite of do_cmd() to treat the optlen parameter consistently and not use it to carry a pointer or integer dependent on the context.
2003-09-04 15:57:37 +00:00
if (do_cmd(ocmd, data, (uintptr_t)&nbytes) < 0)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
err(EX_OSERR, "getsockopt(IP_%s_GET)",
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
co.do_pipe ? "DUMMYNET" : "FW");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
/*
* Count static rules. They have variable size so we
* need to scan the list to count them.
*/
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
for (nstat = 1, r = data, lim = (char *)data + nbytes;
Use IPFW_DEFAULT_RULE instead of hardcoded value since now it is available. MFC after: 5 days.
2008-09-06 17:23:37 +00:00
r->rulenum < IPFW_DEFAULT_RULE && (char *)r < lim;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
++nstat, r = NEXT(r) )
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
; /* nothing */
/*
* Count dynamic rules. This is easier as they have
* fixed size.
*/
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
r = NEXT(r);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
dynrules = (ipfw_dyn_rule *)r ;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
n = (char *)r - (char *)data;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
ndyn = (nbytes - n) / sizeof *dynrules;
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
/* if showing stats, figure out column widths ahead of time */
bcwidth = pcwidth = 0;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
if (show_counters) {
for (n = 0, r = data; n < nstat; n++, r = NEXT(r)) {
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
/* skip rules from another set */
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set && r->set != co.use_set - 1)
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
continue;
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
/* packet counter */
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
width = snprintf(NULL, 0, "%llu",
align_uint64(&r->pcnt));
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
if (width > pcwidth)
pcwidth = width;
/* byte counter */
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
width = snprintf(NULL, 0, "%llu",
align_uint64(&r->bcnt));
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
if (width > bcwidth)
bcwidth = width;
}
}
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.do_dynamic && ndyn) {
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
for (n = 0, d = dynrules; n < ndyn; n++, d++) {
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set) {
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
/* skip rules from another set */
o Fix bug I introduced in the previous commit (ipfw set extention): pack a set number correctly. Submitted by: oleg o Plug a memory leak. Submitted by: oleg and Andrey V. Elsukov Approved by: re (kensmith) MFC after: 1 week
2007-08-26 18:38:31 +00:00
bcopy((char *)&d->rule + sizeof(uint16_t),
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
&set, sizeof(uint8_t));
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (set != co.use_set - 1)
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
continue;
}
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
width = snprintf(NULL, 0, "%llu",
align_uint64(&d->pcnt));
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
if (width > pcwidth)
pcwidth = width;
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
width = snprintf(NULL, 0, "%llu",
align_uint64(&d->bcnt));
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
if (width > bcwidth)
bcwidth = width;
}
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/* if no rule numbers were specified, list all rules */
if (ac == 0) {
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
for (n = 0, r = data; n < nstat; n++, r = NEXT(r)) {
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set && r->set != co.use_set - 1)
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
continue;
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
show_ipfw(r, pcwidth, bcwidth);
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.do_dynamic && ndyn) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
printf("## Dynamic rules (%d):\n", ndyn);
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
for (n = 0, d = dynrules; n < ndyn; n++, d++) {
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set) {
o Fix bug I introduced in the previous commit (ipfw set extention): pack a set number correctly. Submitted by: oleg o Plug a memory leak. Submitted by: oleg and Andrey V. Elsukov Approved by: re (kensmith) MFC after: 1 week
2007-08-26 18:38:31 +00:00
bcopy((char *)&d->rule + sizeof(uint16_t),
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
&set, sizeof(uint8_t));
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (set != co.use_set - 1)
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
continue;
}
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
show_dyn_ipfw(d, pcwidth, bcwidth);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
goto done;
}
/* display specific rules requested on command line */
for (lac = ac, lav = av; lac != 0; lac--) {
/* convert command line rule # */
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
last = rnum = strtoul(*lav++, &endptr, 10);
if (*endptr == '-')
last = strtoul(endptr+1, &endptr, 10);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (*endptr) {
exitval = EX_USAGE;
warnx("invalid rule number: %s", *(lav - 1));
continue;
}
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
for (n = seen = 0, r = data; n < nstat; n++, r = NEXT(r) ) {
if (r->rulenum > last)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set && r->set != co.use_set - 1)
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
continue;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
if (r->rulenum >= rnum && r->rulenum <= last) {
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
show_ipfw(r, pcwidth, bcwidth);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
seen = 1;
}
}
if (!seen) {
/* give precedence to other error(s) */
if (exitval == EX_OK)
exitval = EX_UNAVAILABLE;
warnx("rule %lu does not exist", rnum);
}
}
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.do_dynamic && ndyn) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
printf("## Dynamic rules:\n");
for (lac = ac, lav = av; lac != 0; lac--) {
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
last = rnum = strtoul(*lav++, &endptr, 10);
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
if (*endptr == '-')
last = strtoul(endptr+1, &endptr, 10);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (*endptr)
/* already warned */
continue;
for (n = 0, d = dynrules; n < ndyn; n++, d++) {
Change handling to support strong alignment architectures such as alpha and sparc64. PR: alpha/50658 Submitted by: rizzo Tested on: alpha
2003-06-04 01:17:37 +00:00
uint16_t rulenum;
bcopy(&d->rule, &rulenum, sizeof(rulenum));
if (rulenum > rnum)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set) {
o Fix bug I introduced in the previous commit (ipfw set extention): pack a set number correctly. Submitted by: oleg o Plug a memory leak. Submitted by: oleg and Andrey V. Elsukov Approved by: re (kensmith) MFC after: 1 week
2007-08-26 18:38:31 +00:00
bcopy((char *)&d->rule + sizeof(uint16_t),
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
&set, sizeof(uint8_t));
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (set != co.use_set - 1)
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
continue;
}
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
if (r->rulenum >= rnum && r->rulenum <= last)
o Partially revert rev. 1.103, fix 'ipfw show': dynamically adjust a width of fields for packets and bytes counters. PR: bin/47196 Reviewed by: -audit Not objected by: luigi, des o Use %llu instead of deprecated %qu convert specification for ipfw packets and bytes counters. Noted by: des MFC after: 1 month
2003-03-13 10:51:53 +00:00
show_dyn_ipfw(d, pcwidth, bcwidth);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
}
}
ac = 0;
done:
free(data);
if (exitval != EX_OK)
exit(exitval);
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
#undef NEXT
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
static int
lookup_host (char *host, struct in_addr *ipaddr)
{
struct hostent *he;
if (!inet_aton(host, ipaddr)) {
if ((he = gethostbyname(host)) == NULL)
return(-1);
*ipaddr = *(struct in_addr *)he->h_addr_list[0];
}
return(0);
}
/*
* fills the addr and mask fields in the instruction as appropriate from av.
* Update length as appropriate.
* The following formats are allowed:
* me returns O_IP_*_ME
* 1.2.3.4 single IP address
* 1.2.3.4:5.6.7.8 address:mask
* 1.2.3.4/24 address/mask
* 1.2.3.4/26{1,6,5,4,23} set of addresses in a subnet
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
* We can have multiple comma-separated address/mask entries.
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
static void
fill_ip(ipfw_insn_ip *cmd, char *av)
{
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
int len = 0;
uint32_t *d = ((ipfw_insn_u32 *)cmd)->d;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->o.len &= ~F_LEN_MASK; /* zero len */
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(av, "any") == 0)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return;
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(av, "me") == 0) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->o.len |= F_INSN_SIZE(ipfw_insn);
return;
}
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (strncmp(av, "table(", 6) == 0) {
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
char *p = strchr(av + 6, ',');
if (p)
*p++ = '\0';
cmd->o.opcode = O_IP_DST_LOOKUP;
cmd->o.arg1 = strtoul(av + 6, NULL, 0);
if (p) {
cmd->o.len |= F_INSN_SIZE(ipfw_insn_u32);
d[0] = strtoul(p, NULL, 0);
} else
cmd->o.len |= F_INSN_SIZE(ipfw_insn);
return;
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
while (av) {
/*
* After the address we can have '/' or ':' indicating a mask,
* ',' indicating another address follows, '{' indicating a
* set of addresses of unspecified size.
*/
Fix a parsing bug when specifying more than one address with dotted decimal netmask. Reported by: Igor Anishchuk PR: kern/107565 MFC after: 3 days
2007-01-07 03:02:02 +00:00
char *t = NULL, *p = strpbrk(av, "/:,{");
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
int masklen;
Fix a number of (innocuous) warnings, and remove a useless test. There are still several signed/unsigned warnings left, which require a bit more study for a proper fix. This file has grown beyond reasonable limits. We really need to split it into separate components (ipv4, ipv6, dummynet, nat, table, userland-kernel communication ...) so we can make mainteinance easier. MFC after: 1 weeks
2009-01-20 18:16:31 +00:00
char md, nd = '\0';
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (p) {
md = *p;
*p++ = '\0';
Fix a parsing bug when specifying more than one address with dotted decimal netmask. Reported by: Igor Anishchuk PR: kern/107565 MFC after: 3 days
2007-01-07 03:02:02 +00:00
if ((t = strpbrk(p, ",{")) != NULL) {
nd = *t;
*t = '\0';
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
} else
md = '\0';
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
if (lookup_host(av, (struct in_addr *)&d[0]) != 0)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
errx(EX_NOHOST, "hostname ``%s'' unknown", av);
switch (md) {
case ':':
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
if (!inet_aton(p, (struct in_addr *)&d[1]))
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
errx(EX_DATAERR, "bad netmask ``%s''", p);
break;
case '/':
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
masklen = atoi(p);
if (masklen == 0)
d[1] = htonl(0); /* mask */
else if (masklen > 32)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
errx(EX_DATAERR, "bad width ``%s''", p);
else
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
d[1] = htonl(~0 << (32 - masklen));
break;
case '{': /* no mask, assume /24 and put back the '{' */
d[1] = htonl(~0 << (32 - 24));
*(--p) = md;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
case ',': /* single address plus continuation */
*(--p) = md;
/* FALLTHROUGH */
case 0: /* initialization value */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
default:
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
d[1] = htonl(~0); /* force /32 */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
d[0] &= d[1]; /* mask base address with mask */
Fix a parsing bug when specifying more than one address with dotted decimal netmask. Reported by: Igor Anishchuk PR: kern/107565 MFC after: 3 days
2007-01-07 03:02:02 +00:00
if (t)
*t = nd;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
/* find next separator */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (p)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
p = strpbrk(p, ",{");
if (p && *p == '{') {
/*
* We have a set of addresses. They are stored as follows:
* arg1 is the set size (powers of 2, 2..256)
* addr is the base address IN HOST FORMAT
* mask.. is an array of arg1 bits (rounded up to
* the next multiple of 32) with bits set
* for each host in the map.
*/
uint32_t *map = (uint32_t *)&cmd->mask;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int low, high;
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
int i = contigmask((uint8_t *)&(d[1]), 32);
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
if (len > 0)
errx(EX_DATAERR, "address set cannot be in a list");
if (i < 24 || i > 31)
errx(EX_DATAERR, "invalid set with mask %d\n", i);
cmd->o.arg1 = 1<<(32-i); /* map length */
d[0] = ntohl(d[0]); /* base addr in host format */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->o.opcode = O_IP_DST_SET; /* default */
cmd->o.len |= F_INSN_SIZE(ipfw_insn_u32) + (cmd->o.arg1+31)/32;
Forgot this one: properly initialize an address set when the set size is less than 32 bits (/28 mask or more). Also remove a debugging fprintf().
2002-07-31 22:42:08 +00:00
for (i = 0; i < (cmd->o.arg1+31)/32 ; i++)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
map[i] = 0; /* clear map */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
av = p + 1;
low = d[0] & 0xff;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
high = low + cmd->o.arg1 - 1;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
/*
* Here, i stores the previous value when we specify a range
* of addresses within a mask, e.g. 45-63. i = -1 means we
* have no previous value.
*/
syntactic sugar: support range notation such as 1.2.3.4/24{5,6,7,10-20,60-90} for set of ip addresses. Previously you needed to specify every address in the range, which was unconvenient and lead to very long lines. Internally the set is still stored in the same way, just the input and output routines are modified. Manpage update still missing. Perhaps a similar preprocessing step would be useful for port ranges. MFC after: 3 days
2003-06-23 08:20:28 +00:00
i = -1; /* previous value in a range */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
while (isdigit(*av)) {
char *s;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
int a = strtol(av, &s, 0);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
if (s == av) { /* no parameter */
if (*av != '}')
errx(EX_DATAERR, "set not closed\n");
if (i != -1)
errx(EX_DATAERR, "incomplete range %d-", i);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
if (a < low || a > high)
errx(EX_DATAERR, "addr %d out of range [%d-%d]\n",
a, low, high);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
a -= low;
syntactic sugar: support range notation such as 1.2.3.4/24{5,6,7,10-20,60-90} for set of ip addresses. Previously you needed to specify every address in the range, which was unconvenient and lead to very long lines. Internally the set is still stored in the same way, just the input and output routines are modified. Manpage update still missing. Perhaps a similar preprocessing step would be useful for port ranges. MFC after: 3 days
2003-06-23 08:20:28 +00:00
if (i == -1) /* no previous in range */
i = a;
else { /* check that range is valid */
if (i > a)
errx(EX_DATAERR, "invalid range %d-%d",
i+low, a+low);
if (*s == '-')
errx(EX_DATAERR, "double '-' in range");
}
for (; i <= a; i++)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
map[i/32] |= 1<<(i & 31);
syntactic sugar: support range notation such as 1.2.3.4/24{5,6,7,10-20,60-90} for set of ip addresses. Previously you needed to specify every address in the range, which was unconvenient and lead to very long lines. Internally the set is still stored in the same way, just the input and output routines are modified. Manpage update still missing. Perhaps a similar preprocessing step would be useful for port ranges. MFC after: 3 days
2003-06-23 08:20:28 +00:00
i = -1;
if (*s == '-')
i = a;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
else if (*s == '}')
syntactic sugar: support range notation such as 1.2.3.4/24{5,6,7,10-20,60-90} for set of ip addresses. Previously you needed to specify every address in the range, which was unconvenient and lead to very long lines. Internally the set is still stored in the same way, just the input and output routines are modified. Manpage update still missing. Perhaps a similar preprocessing step would be useful for port ranges. MFC after: 3 days
2003-06-23 08:20:28 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
av = s+1;
}
return;
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
av = p;
if (av) /* then *av must be a ',' */
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
/* Check this entry */
if (d[1] == 0) { /* "any", specified as x.x.x.x/0 */
/*
* 'any' turns the entire list into a NOP.
* 'not any' never matches, so it is removed from the
* list unless it is the only item, in which case we
* report an error.
*/
if (cmd->o.len & F_NOT) { /* "not any" never matches */
if (av == NULL && len == 0) /* only this entry */
errx(EX_DATAERR, "not any never matches");
}
/* else do nothing and skip this entry */
o Fix an incorrect parsing of 0.0.0.0/0 expression. PR: kern/64778 MFC after: 6 weeks
2004-04-09 17:26:01 +00:00
return;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
}
/* A single IP can be stored in an optimized format */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (d[1] == (uint32_t)~0 && av == NULL && len == 0) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->o.len |= F_INSN_SIZE(ipfw_insn_u32);
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
return;
}
len += 2; /* two words... */
d += 2;
} /* end while */
Check the length of the ipv4 and ipv6 address lists. It must be less than F_LEN_MASK. MFC after: 5 days
2006-09-16 19:27:40 +00:00
if (len + 1 > F_LEN_MASK)
errx(EX_DATAERR, "address list too long");
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
cmd->o.len |= len+1;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
/* n2mask sets n bits of the mask */
Put dummynet-related code in a separate file. To this purpose, add prototypes for global functions in ipfw2.h and move there also the list of tokens used in various places in the code.
2009-01-27 11:06:59 +00:00
void
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
n2mask(struct in6_addr *mask, int n)
{
static int minimask[9] =
{ 0x00, 0x80, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc, 0xfe, 0xff };
u_char *p;
memset(mask, 0, sizeof(struct in6_addr));
p = (u_char *) mask;
for (; n > 0; p++, n -= 8) {
if (n >= 8)
*p = 0xff;
else
*p = minimask[n];
}
return;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* helper function to process a set of flags and set bits in the
* appropriate masks.
*/
static void
fill_flags(ipfw_insn *cmd, enum ipfw_opcodes opcode,
struct _s_x *flags, char *p)
{
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint8_t set=0, clear=0;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
while (p && *p) {
char *q; /* points to the separator */
int val;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint8_t *which; /* mask we are working on */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (*p == '!') {
p++;
which = &clear;
} else
which = &set;
q = strchr(p, ',');
if (q)
*q++ = '\0';
val = match_token(flags, p);
if (val <= 0)
errx(EX_DATAERR, "invalid flag %s", p);
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
*which |= (uint8_t)val;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
p = q;
}
cmd->opcode = opcode;
cmd->len = (cmd->len & (F_NOT | F_OR)) | 1;
cmd->arg1 = (set & 0xff) | ( (clear & 0xff) << 8);
}
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
void
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
ipfw_delete(char *av[])
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
uint32_t rulenum;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int i;
int exitval = EX_OK;
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
int do_set = 0;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
o Move NEED1 macro to the top of the source file. o Add sanity checking to the firewall delete operation which tells the user that a firewall rule specification is required. The previous behaviour was to exit without reporting any errors to the user. Approved by: bmilekic (mentor)
2004-06-02 21:12:55 +00:00
NEED1("missing rule specification");
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if ( *av && _substrcmp(*av, "set") == 0) {
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
/* Do not allow using the following syntax:
* ipfw set N delete set M
*/
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set)
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
errx(EX_DATAERR, "invalid syntax");
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
do_set = 1; /* delete set */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/* Rule number */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
while (*av && isdigit(**av)) {
i = atoi(*av); av++;
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.do_nat) {
Summer of Code 2005: improve libalias - part 2 of 2 With the second (and last) part of my previous Summer of Code work, we get: -ipfw's in kernel nat -redirect_* and LSNAT support General information about nat syntax and some examples are available in the ipfw (8) man page. The redirect and LSNAT syntax are identical to natd, so please refer to natd (8) man page. To enable in kernel nat in rc.conf, two options were added: o firewall_nat_enable: equivalent to natd_enable o firewall_nat_interface: equivalent to natd_interface Remember to set net.inet.ip.fw.one_pass to 0, if you want the packet to continue being checked by the firewall ruleset after being (de)aliased. NOTA BENE: due to some problems with libalias architecture, in kernel nat won't work with TSO enabled nic, thus you have to disable TSO via ifconfig (ifconfig foo0 -tso). Approved by: glebius (mentor)
2006-12-29 21:59:17 +00:00
exitval = do_cmd(IP_FW_NAT_DEL, &i, sizeof i);
if (exitval) {
exitval = EX_UNAVAILABLE;
warn("rule %u not available", i);
}
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
} else if (co.do_pipe) {
Put dummynet-related code in a separate file. To this purpose, add prototypes for global functions in ipfw2.h and move there also the list of tokens used in various places in the code.
2009-01-27 11:06:59 +00:00
exitval = ipfw_delete_pipe(co.do_pipe, i);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
} else {
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set)
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
rulenum = (i & 0xffff) | (5 << 24) |
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
((co.use_set - 1) << 16);
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
else
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
rulenum = (i & 0xffff) | (do_set << 24);
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
i = do_cmd(IP_FW_DEL, &rulenum, sizeof rulenum);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (i) {
exitval = EX_UNAVAILABLE;
warn("rule %u: setsockopt(IP_FW_DEL)",
rulenum);
}
}
}
if (exitval != EX_OK)
exit(exitval);
}
/*
* fill the interface structure. We do not check the name as we can
* create interfaces dynamically, so checking them at insert time
* makes relatively little sense.
Replace the if_name and if_unit members of struct ifnet with new members if_xname, if_dname, and if_dunit. if_xname is the name of the interface and if_dname/unit are the driver name and instance. This change paves the way for interface renaming and enhanced pseudo device creation and configuration symantics. Approved By: re (in principle) Reviewed By: njl, imp Tested On: i386, amd64, sparc64 Obtained From: NetBSD (if_xname)
2003-10-31 18:32:15 +00:00
* Interface names containing '*', '?', or '[' are assumed to be shell
* patterns which match interfaces.
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
static void
fill_iface(ipfw_insn_if *cmd, char *arg)
{
cmd->name[0] = '\0';
cmd->o.len |= F_INSN_SIZE(ipfw_insn_if);
/* Parse the interface or address */
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (strcmp(arg, "any") == 0)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->o.len = 0; /* effectively ignore this command */
else if (!isdigit(*arg)) {
Replace the if_name and if_unit members of struct ifnet with new members if_xname, if_dname, and if_dunit. if_xname is the name of the interface and if_dname/unit are the driver name and instance. This change paves the way for interface renaming and enhanced pseudo device creation and configuration symantics. Approved By: re (in principle) Reviewed By: njl, imp Tested On: i386, amd64, sparc64 Obtained From: NetBSD (if_xname)
2003-10-31 18:32:15 +00:00
strlcpy(cmd->name, arg, sizeof(cmd->name));
cmd->p.glob = strpbrk(arg, "*?[") != NULL ? 1 : 0;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
} else if (!inet_aton(arg, &cmd->p.ip))
errx(EX_DATAERR, "bad ip address ``%s''", arg);
}
static void
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
get_mac_addr_mask(const char *p, uint8_t *addr, uint8_t *mask)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
int i;
size_t l;
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
char *ap, *ptr, *optr;
struct ether_addr *mac;
const char *macset = "0123456789abcdefABCDEF:";
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
if (strcmp(p, "any") == 0) {
for (i = 0; i < ETHER_ADDR_LEN; i++)
addr[i] = mask[i] = 0;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return;
}
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
optr = ptr = strdup(p);
if ((ap = strsep(&ptr, "&/")) != NULL && *ap != 0) {
l = strlen(ap);
if (strspn(ap, macset) != l || (mac = ether_aton(ap)) == NULL)
errx(EX_DATAERR, "Incorrect MAC address");
bcopy(mac, addr, ETHER_ADDR_LEN);
} else
errx(EX_DATAERR, "Incorrect MAC address");
if (ptr != NULL) { /* we have mask? */
if (p[ptr - optr - 1] == '/') { /* mask len */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
long ml = strtol(ptr, &ap, 10);
if (*ap != 0 || ml > ETHER_ADDR_LEN * 8 || ml < 0)
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
errx(EX_DATAERR, "Incorrect mask length");
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
for (i = 0; ml > 0 && i < ETHER_ADDR_LEN; ml -= 8, i++)
mask[i] = (ml >= 8) ? 0xff: (~0) << (8 - ml);
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
} else { /* mask */
l = strlen(ptr);
if (strspn(ptr, macset) != l ||
(mac = ether_aton(ptr)) == NULL)
errx(EX_DATAERR, "Incorrect mask");
bcopy(mac, mask, ETHER_ADDR_LEN);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
} else { /* default mask: ff:ff:ff:ff:ff:ff */
for (i = 0; i < ETHER_ADDR_LEN; i++)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
mask[i] = 0xff;
}
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
for (i = 0; i < ETHER_ADDR_LEN; i++)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
addr[i] &= mask[i];
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
free(optr);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
/*
* helper function, updates the pointer to cmd with the length
* of the current command, and also cleans up the first word of
* the new command in case it has been clobbered before.
*/
static ipfw_insn *
next_cmd(ipfw_insn *cmd)
{
cmd += F_LEN(cmd);
bzero(cmd, sizeof(*cmd));
return cmd;
}
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
/*
* Takes arguments and copies them into a comment
*/
static void
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
fill_comment(ipfw_insn *cmd, char **av)
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
{
int i, l;
char *p = (char *)(cmd + 1);
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
cmd->opcode = O_NOP;
cmd->len = (cmd->len & (F_NOT | F_OR));
/* Compute length of comment string. */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
for (i = 0, l = 0; av[i] != NULL; i++)
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
l += strlen(av[i]) + 1;
if (l == 0)
return;
if (l > 84)
errx(EX_DATAERR,
"comment too long (max 80 chars)");
l = 1 + (l+3)/4;
cmd->len = (cmd->len & (F_NOT | F_OR)) | l;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
for (i = 0; av[i] != NULL; i++) {
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
strcpy(p, av[i]);
p += strlen(av[i]);
*p++ = ' ';
}
*(--p) = '\0';
}
Fix one typo in help() string, remove whitespace at end of line and other minor whitespace changes. Replace u_char with uint8_t in a few places.
2003-07-14 18:57:41 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* A function to fill simple commands of size 1.
* Existing flags are preserved.
*/
static void
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
fill_cmd(ipfw_insn *cmd, enum ipfw_opcodes opcode, int flags, uint16_t arg)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
cmd->opcode = opcode;
cmd->len = ((cmd->len | flags) & (F_NOT | F_OR)) | 1;
cmd->arg1 = arg;
}
/*
* Fetch and add the MAC address and type, with masks. This generates one or
* two microinstructions, and returns the pointer to the last one.
*/
static ipfw_insn *
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
add_mac(ipfw_insn *cmd, char *av[])
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
ipfw_insn_mac *mac;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if ( ( av[0] == NULL ) || ( av[1] == NULL ) )
One more (hopefully the last one) step in cleaning up the syntax, following Julian's good suggestion: since you can specify any match pattern as an option, rules now have the following format: [<proto> from <src> to <dst>] [options] i.e. the first part is now entirely optional (and left there just for compatibility with ipfw1 rulesets). Add a "-c" flag to show/list rules in the compact form (i.e. without the "ip from any to any" part) when possible. The default is to include it so that scripts processing ipfw's canonical output will still work. Note that as part of this cleanup (and to remove ambiguity), MAC fields now can only be specified in the options part. Update the manpage to reflect the syntax. Clarify the behaviour when a match is attempted on fields which are not present in the packet, e.g. port numbers on non TCP/UDP packets, and the "not" operator is specified. E.g. ipfw add allow not src-port 80 will match also ICMP packets because they do not have port numbers, so "src-port 80" will fail and "not src-port 80" will succeed. For such cases it is advised to insert further options to prevent undesired results (e.g. in the case above, "ipfw add allow proto tcp not src-port 80"). We definitely need to rewrite the parser using lex and yacc!
2002-08-19 12:36:54 +00:00
errx(EX_DATAERR, "MAC dst src");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->opcode = O_MACADDR2;
cmd->len = (cmd->len & (F_NOT | F_OR)) | F_INSN_SIZE(ipfw_insn_mac);
mac = (ipfw_insn_mac *)cmd;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
get_mac_addr_mask(av[0], mac->addr, mac->mask); /* dst */
o Teach get_mac_addr_mask() to not silently accept incorrect MAC addresses. o Swap a couple of magic 6s by ETHER_ADDR_LEN. PR: bin/80913 Submitted by: Andrey V. Elsukov MFC after: 1 month
2007-05-09 18:31:49 +00:00
get_mac_addr_mask(av[1], &(mac->addr[ETHER_ADDR_LEN]),
&(mac->mask[ETHER_ADDR_LEN])); /* src */
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
return cmd;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
static ipfw_insn *
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
add_mactype(ipfw_insn *cmd, char *av)
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
{
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (!av)
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
errx(EX_DATAERR, "missing MAC type");
if (strcmp(av, "any") != 0) { /* we have a non-null type */
fill_newports((ipfw_insn_u16 *)cmd, av, IPPROTO_ETHERTYPE);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->opcode = O_MAC_TYPE;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
return cmd;
} else
return NULL;
}
static ipfw_insn *
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
add_proto0(ipfw_insn *cmd, char *av, u_char *protop)
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
{
struct protoent *pe;
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
char *ep;
int proto;
Revert `proto ip' back to the previous behavior. The kernel side of ipfw2 doesn't allow zero as protocol number. MFC after: 3 days
2006-03-05 15:55:46 +00:00
proto = strtol(av, &ep, 10);
if (*ep != '\0' || proto <= 0) {
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
if ((pe = getprotobyname(av)) == NULL)
return NULL;
proto = pe->p_proto;
}
fill_cmd(cmd, O_PROTO, 0, proto);
*protop = proto;
return cmd;
}
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
static ipfw_insn *
add_proto(ipfw_insn *cmd, char *av, u_char *protop)
{
u_char proto = IPPROTO_IP;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
Revert `proto ip' back to the previous behavior. The kernel side of ipfw2 doesn't allow zero as protocol number. MFC after: 3 days
2006-03-05 15:55:46 +00:00
if (_substrcmp(av, "all") == 0 || strcmp(av, "ip") == 0)
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
; /* do not set O_IP4 nor O_IP6 */
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
else if (strcmp(av, "ip4") == 0)
/* explicit "just IPv4" rule */
fill_cmd(cmd, O_IP4, 0, 0);
else if (strcmp(av, "ip6") == 0) {
/* explicit "just IPv6" rule */
proto = IPPROTO_IPV6;
fill_cmd(cmd, O_IP6, 0, 0);
} else
return add_proto0(cmd, av, protop);
*protop = proto;
return cmd;
}
static ipfw_insn *
add_proto_compat(ipfw_insn *cmd, char *av, u_char *protop)
{
u_char proto = IPPROTO_IP;
if (_substrcmp(av, "all") == 0 || strcmp(av, "ip") == 0)
; /* do not set O_IP4 nor O_IP6 */
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
else if (strcmp(av, "ipv4") == 0 || strcmp(av, "ip4") == 0)
/* explicit "just IPv4" rule */
fill_cmd(cmd, O_IP4, 0, 0);
else if (strcmp(av, "ipv6") == 0 || strcmp(av, "ip6") == 0) {
/* explicit "just IPv6" rule */
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
proto = IPPROTO_IPV6;
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
fill_cmd(cmd, O_IP6, 0, 0);
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
} else
return add_proto0(cmd, av, protop);
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
*protop = proto;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
return cmd;
}
static ipfw_insn *
add_srcip(ipfw_insn *cmd, char *av)
{
fill_ip((ipfw_insn_ip *)cmd, av);
if (cmd->opcode == O_IP_DST_SET) /* set */
cmd->opcode = O_IP_SRC_SET;
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
else if (cmd->opcode == O_IP_DST_LOOKUP) /* table */
cmd->opcode = O_IP_SRC_LOOKUP;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
else if (F_LEN(cmd) == F_INSN_SIZE(ipfw_insn)) /* me */
cmd->opcode = O_IP_SRC_ME;
else if (F_LEN(cmd) == F_INSN_SIZE(ipfw_insn_u32)) /* one IP */
cmd->opcode = O_IP_SRC;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
else /* addr/mask */
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
cmd->opcode = O_IP_SRC_MASK;
return cmd;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
static ipfw_insn *
add_dstip(ipfw_insn *cmd, char *av)
{
fill_ip((ipfw_insn_ip *)cmd, av);
if (cmd->opcode == O_IP_DST_SET) /* set */
;
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
else if (cmd->opcode == O_IP_DST_LOOKUP) /* table */
;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
else if (F_LEN(cmd) == F_INSN_SIZE(ipfw_insn)) /* me */
cmd->opcode = O_IP_DST_ME;
else if (F_LEN(cmd) == F_INSN_SIZE(ipfw_insn_u32)) /* one IP */
cmd->opcode = O_IP_DST;
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
else /* addr/mask */
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
cmd->opcode = O_IP_DST_MASK;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return cmd;
}
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
static ipfw_insn *
add_ports(ipfw_insn *cmd, char *av, u_char proto, int opcode)
{
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
/* XXX "any" is trapped before. Perhaps "to" */
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(av, "any") == 0) {
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
return NULL;
} else if (fill_newports((ipfw_insn_u16 *)cmd, av, proto)) {
/* XXX todo: check that we have a protocol with ports */
cmd->opcode = opcode;
return cmd;
}
return NULL;
}
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
static ipfw_insn *
add_src(ipfw_insn *cmd, char *av, u_char proto)
{
struct in6_addr a;
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
char *host, *ch;
ipfw_insn *ret = NULL;
if ((host = strdup(av)) == NULL)
return NULL;
if ((ch = strrchr(host, '/')) != NULL)
*ch = '\0';
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
if (proto == IPPROTO_IPV6 || strcmp(av, "me6") == 0 ||
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
inet_pton(AF_INET6, host, &a) == 1)
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
ret = add_srcip6(cmd, av);
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
/* XXX: should check for IPv4, not !IPv6 */
Regigle parens to try and get the intended affect. This should fix people having trouble with the "me6" keyword. Also, we were using inet_pton on the wrong variable in one place. Reviewed by: mlaier (previous version of patch) Obtained from: Sascha Blank (inet_pton change) MFC after: 1 week
2006-08-20 20:10:36 +00:00
if (ret == NULL && (proto == IPPROTO_IP || strcmp(av, "me") == 0 ||
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
inet_pton(AF_INET6, host, &a) != 1))
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
ret = add_srcip(cmd, av);
Regigle parens to try and get the intended affect. This should fix people having trouble with the "me6" keyword. Also, we were using inet_pton on the wrong variable in one place. Reviewed by: mlaier (previous version of patch) Obtained from: Sascha Blank (inet_pton change) MFC after: 1 week
2006-08-20 20:10:36 +00:00
if (ret == NULL && strcmp(av, "any") != 0)
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
ret = cmd;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
free(host);
return ret;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
}
static ipfw_insn *
add_dst(ipfw_insn *cmd, char *av, u_char proto)
{
struct in6_addr a;
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
char *host, *ch;
ipfw_insn *ret = NULL;
if ((host = strdup(av)) == NULL)
return NULL;
if ((ch = strrchr(host, '/')) != NULL)
*ch = '\0';
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
if (proto == IPPROTO_IPV6 || strcmp(av, "me6") == 0 ||
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
inet_pton(AF_INET6, host, &a) == 1)
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
ret = add_dstip6(cmd, av);
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
/* XXX: should check for IPv4, not !IPv6 */
Regigle parens to try and get the intended affect. This should fix people having trouble with the "me6" keyword. Also, we were using inet_pton on the wrong variable in one place. Reviewed by: mlaier (previous version of patch) Obtained from: Sascha Blank (inet_pton change) MFC after: 1 week
2006-08-20 20:10:36 +00:00
if (ret == NULL && (proto == IPPROTO_IP || strcmp(av, "me") == 0 ||
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
inet_pton(AF_INET6, host, &a) != 1))
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
ret = add_dstip(cmd, av);
Regigle parens to try and get the intended affect. This should fix people having trouble with the "me6" keyword. Also, we were using inet_pton on the wrong variable in one place. Reviewed by: mlaier (previous version of patch) Obtained from: Sascha Blank (inet_pton change) MFC after: 1 week
2006-08-20 20:10:36 +00:00
if (ret == NULL && strcmp(av, "any") != 0)
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
ret = cmd;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
For src/dest parsing take off the netmask before checking for AF with inet_pton. This fixes cases like "fe02::/16". PR: bin/91245 Reported by: Fredrik Lindberge
2006-05-14 03:53:04 +00:00
free(host);
return ret;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* Parse arguments and assemble the microinstructions which make up a rule.
* Rules are added into the 'rulebuf' and then copied in the correct order
* into the actual rule.
*
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
* The syntax for a rule starts with the action, followed by
* optional action parameters, and the various match patterns.
Correct typos, mostly s/ a / an / where appropriate. Some whitespace cleanup, especially in troff files.
2003-01-01 18:49:04 +00:00
* In the assembled microcode, the first opcode must be an O_PROBE_STATE
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* (generated if the rule includes a keep-state option), then the
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
* various match patterns, log/altq actions, and the actual action.
Kill EOL whitespaces, style(9) fix.
2002-11-06 15:09:34 +00:00
*
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
void
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
ipfw_add(char *av[])
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
/*
* rules are added into the 'rulebuf' and then copied in
* the correct order into the actual rule.
* Some things that need to go out of order (prob, action etc.)
* go into actbuf[].
*/
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
static uint32_t rulebuf[255], actbuf[255], cmdbuf[255];
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
ipfw_insn *src, *dst, *cmd, *action, *prev=NULL;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
ipfw_insn *first_cmd; /* first match pattern */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
struct ip_fw *rule;
/*
* various flags used to record that we entered some fields.
*/
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
ipfw_insn *have_state = NULL; /* check-state or keep-state */
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
ipfw_insn *have_log = NULL, *have_altq = NULL, *have_tag = NULL;
o Restore a historical ipfw1 logamount behaviour: rules with 'log' keyword but without 'logamount' limit the amount of their log messages by net.inet.ip.fw.verbose_limit sysctl value. RELENG_5 candidate. PR: kern/46080 Submitted by: Dan Pelleg MFC after: 1 week
2004-08-29 08:25:02 +00:00
size_t len;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int i;
int open_par = 0; /* open parenthesis ( */
/* proto is here because it is used to fetch ports */
u_char proto = IPPROTO_IP; /* default protocol */
Fix a kernel panic with rules of the type prob 0.5 pipe NN .... due to the generation of an invalid ipfw instruction sequence. No ABI change, but you need to upgrade /sbin/ipfw to generate the correct code. Approved by: re
2002-11-26 19:58:12 +00:00
double match_prob = 1; /* match probability, default is always match */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
bzero(actbuf, sizeof(actbuf)); /* actions go here */
bzero(cmdbuf, sizeof(cmdbuf));
bzero(rulebuf, sizeof(rulebuf));
rule = (struct ip_fw *)rulebuf;
cmd = (ipfw_insn *)cmdbuf;
action = (ipfw_insn *)actbuf;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/* [rule N] -- Rule number optional */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && isdigit(**av)) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
rule->rulenum = atoi(*av);
av++;
}
Userland side of: Allow set 31 to be used for rules other than 65535. Set 31 is still special because rules belonging to it are not deleted by the "ipfw flush" command, but must be deleted explicitly with "ipfw delete set 31" or by individual rule numbers. This implement a flexible form of "persistent rules" which you might want to have available even after an "ipfw flush". Note that this change does not violate POLA, because you could not use set 31 in a ruleset before this change. Suggested by: Paul Richards
2003-07-15 23:08:44 +00:00
/* [set N] -- set number (0..RESVD_SET), optional */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && !av[1] && _substrcmp(*av, "set") == 0) {
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
int set = strtoul(av[1], NULL, 10);
Userland side of: Allow set 31 to be used for rules other than 65535. Set 31 is still special because rules belonging to it are not deleted by the "ipfw flush" command, but must be deleted explicitly with "ipfw delete set 31" or by individual rule numbers. This implement a flexible form of "persistent rules" which you might want to have available even after an "ipfw flush". Note that this change does not violate POLA, because you could not use set 31 in a ruleset before this change. Suggested by: Paul Richards
2003-07-15 23:08:44 +00:00
if (set < 0 || set > RESVD_SET)
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
errx(EX_DATAERR, "illegal set %s", av[1]);
rule->set = set;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av += 2;
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/* [prob D] -- match probability, optional */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && av[1] && _substrcmp(*av, "prob") == 0) {
Fix a kernel panic with rules of the type prob 0.5 pipe NN .... due to the generation of an invalid ipfw instruction sequence. No ABI change, but you need to upgrade /sbin/ipfw to generate the correct code. Approved by: re
2002-11-26 19:58:12 +00:00
match_prob = strtod(av[1], NULL);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Fix a kernel panic with rules of the type prob 0.5 pipe NN .... due to the generation of an invalid ipfw instruction sequence. No ABI change, but you need to upgrade /sbin/ipfw to generate the correct code. Approved by: re
2002-11-26 19:58:12 +00:00
if (match_prob <= 0 || match_prob > 1)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
errx(EX_DATAERR, "illegal match prob. %s", av[1]);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av += 2;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
/* action -- mandatory */
NEED1("missing action");
i = match_token(rule_actions, *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
action->len = 1; /* default */
switch(i) {
case TOK_CHECKSTATE:
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
have_state = action;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
action->opcode = O_CHECK_STATE;
break;
case TOK_ACCEPT:
action->opcode = O_ACCEPT;
break;
case TOK_DENY:
action->opcode = O_DENY;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
action->arg1 = 0;
break;
case TOK_REJECT:
action->opcode = O_REJECT;
action->arg1 = ICMP_UNREACH_HOST;
break;
case TOK_RESET:
action->opcode = O_REJECT;
action->arg1 = ICMP_REJECT_RST;
break;
* Add dynamic sysctl for net.inet6.ip6.fw. * Correct handling of IPv6 Extension Headers. * Add unreach6 code. * Add logging for IPv6. Submitted by: sysctl handling derived from patch from ume needed for ip6fw Obtained from: is_icmp6_query and send_reject6 derived from similar functions of netinet6,ip6fw Reviewed by: ume, gnn; silence on ipfw@ Test setup provided by: CK Software GmbH MFC after: 6 days
2005-08-13 11:02:34 +00:00
case TOK_RESET6:
action->opcode = O_UNREACH6;
action->arg1 = ICMP6_UNREACH_RST;
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case TOK_UNREACH:
action->opcode = O_REJECT;
NEED1("missing reject code");
fill_reject_code(&action->arg1, *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
* Add dynamic sysctl for net.inet6.ip6.fw. * Correct handling of IPv6 Extension Headers. * Add unreach6 code. * Add logging for IPv6. Submitted by: sysctl handling derived from patch from ume needed for ip6fw Obtained from: is_icmp6_query and send_reject6 derived from similar functions of netinet6,ip6fw Reviewed by: ume, gnn; silence on ipfw@ Test setup provided by: CK Software GmbH MFC after: 6 days
2005-08-13 11:02:34 +00:00
case TOK_UNREACH6:
action->opcode = O_UNREACH6;
NEED1("missing unreach code");
fill_unreach6_code(&action->arg1, *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
* Add dynamic sysctl for net.inet6.ip6.fw. * Correct handling of IPv6 Extension Headers. * Add unreach6 code. * Add logging for IPv6. Submitted by: sysctl handling derived from patch from ume needed for ip6fw Obtained from: is_icmp6_query and send_reject6 derived from similar functions of netinet6,ip6fw Reviewed by: ume, gnn; silence on ipfw@ Test setup provided by: CK Software GmbH MFC after: 6 days
2005-08-13 11:02:34 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_COUNT:
action->opcode = O_COUNT;
break;
Add table/tablearg support to ipfw's nat. MFC After: 1 week
2008-02-24 15:37:45 +00:00
case TOK_NAT:
action->opcode = O_NAT;
action->len = F_INSN_SIZE(ipfw_insn_nat);
goto chkarg;
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_QUEUE:
Add a new feature for optimizining ipfw rulesets - substitution of the action argument with the value obtained from table lookup. The feature is now applicable only to "pipe", "queue", "divert", "tee", "netgraph" and "ngtee" rules. An example usage: ipfw pipe 1000 config bw 1000Kbyte/s ipfw pipe 4000 config bw 4000Kbyte/s ipfw table 1 add x.x.x.x 1000 ipfw table 1 add x.x.x.y 4000 ipfw pipe tablearg ip from table(1) to any In the example above the rule will throw different packets to different pipes. TODO: - Support "skipto" action, but without searching all rules. - Improve parser, so that it warns about bad rules. These are: - "tablearg" argument to action, but no "table" in the rule. All traffic will be blocked. - "tablearg" argument to action, but "table" searches for entry with a specific value. All traffic will be blocked. - "tablearg" argument to action, and two "table" looks - for src and for dst. The last lookup will match.
2005-12-13 12:16:03 +00:00
action->opcode = O_QUEUE;
goto chkarg;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_PIPE:
Add a new feature for optimizining ipfw rulesets - substitution of the action argument with the value obtained from table lookup. The feature is now applicable only to "pipe", "queue", "divert", "tee", "netgraph" and "ngtee" rules. An example usage: ipfw pipe 1000 config bw 1000Kbyte/s ipfw pipe 4000 config bw 4000Kbyte/s ipfw table 1 add x.x.x.x 1000 ipfw table 1 add x.x.x.y 4000 ipfw pipe tablearg ip from table(1) to any In the example above the rule will throw different packets to different pipes. TODO: - Support "skipto" action, but without searching all rules. - Improve parser, so that it warns about bad rules. These are: - "tablearg" argument to action, but no "table" in the rule. All traffic will be blocked. - "tablearg" argument to action, but "table" searches for entry with a specific value. All traffic will be blocked. - "tablearg" argument to action, and two "table" looks - for src and for dst. The last lookup will match.
2005-12-13 12:16:03 +00:00
action->opcode = O_PIPE;
goto chkarg;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_SKIPTO:
Add a new feature for optimizining ipfw rulesets - substitution of the action argument with the value obtained from table lookup. The feature is now applicable only to "pipe", "queue", "divert", "tee", "netgraph" and "ngtee" rules. An example usage: ipfw pipe 1000 config bw 1000Kbyte/s ipfw pipe 4000 config bw 4000Kbyte/s ipfw table 1 add x.x.x.x 1000 ipfw table 1 add x.x.x.y 4000 ipfw pipe tablearg ip from table(1) to any In the example above the rule will throw different packets to different pipes. TODO: - Support "skipto" action, but without searching all rules. - Improve parser, so that it warns about bad rules. These are: - "tablearg" argument to action, but no "table" in the rule. All traffic will be blocked. - "tablearg" argument to action, but "table" searches for entry with a specific value. All traffic will be blocked. - "tablearg" argument to action, and two "table" looks - for src and for dst. The last lookup will match.
2005-12-13 12:16:03 +00:00
action->opcode = O_SKIPTO;
goto chkarg;
case TOK_NETGRAPH:
action->opcode = O_NETGRAPH;
goto chkarg;
case TOK_NGTEE:
action->opcode = O_NGTEE;
goto chkarg;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_DIVERT:
Add a new feature for optimizining ipfw rulesets - substitution of the action argument with the value obtained from table lookup. The feature is now applicable only to "pipe", "queue", "divert", "tee", "netgraph" and "ngtee" rules. An example usage: ipfw pipe 1000 config bw 1000Kbyte/s ipfw pipe 4000 config bw 4000Kbyte/s ipfw table 1 add x.x.x.x 1000 ipfw table 1 add x.x.x.y 4000 ipfw pipe tablearg ip from table(1) to any In the example above the rule will throw different packets to different pipes. TODO: - Support "skipto" action, but without searching all rules. - Improve parser, so that it warns about bad rules. These are: - "tablearg" argument to action, but no "table" in the rule. All traffic will be blocked. - "tablearg" argument to action, but "table" searches for entry with a specific value. All traffic will be blocked. - "tablearg" argument to action, and two "table" looks - for src and for dst. The last lookup will match.
2005-12-13 12:16:03 +00:00
action->opcode = O_DIVERT;
goto chkarg;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_TEE:
Add a new feature for optimizining ipfw rulesets - substitution of the action argument with the value obtained from table lookup. The feature is now applicable only to "pipe", "queue", "divert", "tee", "netgraph" and "ngtee" rules. An example usage: ipfw pipe 1000 config bw 1000Kbyte/s ipfw pipe 4000 config bw 4000Kbyte/s ipfw table 1 add x.x.x.x 1000 ipfw table 1 add x.x.x.y 4000 ipfw pipe tablearg ip from table(1) to any In the example above the rule will throw different packets to different pipes. TODO: - Support "skipto" action, but without searching all rules. - Improve parser, so that it warns about bad rules. These are: - "tablearg" argument to action, but no "table" in the rule. All traffic will be blocked. - "tablearg" argument to action, but "table" searches for entry with a specific value. All traffic will be blocked. - "tablearg" argument to action, and two "table" looks - for src and for dst. The last lookup will match.
2005-12-13 12:16:03 +00:00
action->opcode = O_TEE;
chkarg:
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (!av[0])
Add a new feature for optimizining ipfw rulesets - substitution of the action argument with the value obtained from table lookup. The feature is now applicable only to "pipe", "queue", "divert", "tee", "netgraph" and "ngtee" rules. An example usage: ipfw pipe 1000 config bw 1000Kbyte/s ipfw pipe 4000 config bw 4000Kbyte/s ipfw table 1 add x.x.x.x 1000 ipfw table 1 add x.x.x.y 4000 ipfw pipe tablearg ip from table(1) to any In the example above the rule will throw different packets to different pipes. TODO: - Support "skipto" action, but without searching all rules. - Improve parser, so that it warns about bad rules. These are: - "tablearg" argument to action, but no "table" in the rule. All traffic will be blocked. - "tablearg" argument to action, but "table" searches for entry with a specific value. All traffic will be blocked. - "tablearg" argument to action, and two "table" looks - for src and for dst. The last lookup will match.
2005-12-13 12:16:03 +00:00
errx(EX_USAGE, "missing argument for %s", *(av - 1));
if (isdigit(**av)) {
action->arg1 = strtoul(*av, NULL, 10);
if (action->arg1 <= 0 || action->arg1 >= IP_FW_TABLEARG)
errx(EX_DATAERR, "illegal argument for %s",
*(av - 1));
remove a couple of rarely used #define; change PRINT_UINT from a macro to a function (renaming is postponed to reduce clutter)
2009-01-27 07:40:16 +00:00
} else if (_substrcmp(*av, "tablearg") == 0) {
Add a new feature for optimizining ipfw rulesets - substitution of the action argument with the value obtained from table lookup. The feature is now applicable only to "pipe", "queue", "divert", "tee", "netgraph" and "ngtee" rules. An example usage: ipfw pipe 1000 config bw 1000Kbyte/s ipfw pipe 4000 config bw 4000Kbyte/s ipfw table 1 add x.x.x.x 1000 ipfw table 1 add x.x.x.y 4000 ipfw pipe tablearg ip from table(1) to any In the example above the rule will throw different packets to different pipes. TODO: - Support "skipto" action, but without searching all rules. - Improve parser, so that it warns about bad rules. These are: - "tablearg" argument to action, but no "table" in the rule. All traffic will be blocked. - "tablearg" argument to action, but "table" searches for entry with a specific value. All traffic will be blocked. - "tablearg" argument to action, and two "table" looks - for src and for dst. The last lookup will match.
2005-12-13 12:16:03 +00:00
action->arg1 = IP_FW_TABLEARG;
} else if (i == TOK_DIVERT || i == TOK_TEE) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
struct servent *s;
setservent(1);
s = getservbyname(av[0], "divert");
if (s != NULL)
action->arg1 = ntohs(s->s_port);
else
errx(EX_DATAERR, "illegal divert/tee port");
Add a new feature for optimizining ipfw rulesets - substitution of the action argument with the value obtained from table lookup. The feature is now applicable only to "pipe", "queue", "divert", "tee", "netgraph" and "ngtee" rules. An example usage: ipfw pipe 1000 config bw 1000Kbyte/s ipfw pipe 4000 config bw 4000Kbyte/s ipfw table 1 add x.x.x.x 1000 ipfw table 1 add x.x.x.y 4000 ipfw pipe tablearg ip from table(1) to any In the example above the rule will throw different packets to different pipes. TODO: - Support "skipto" action, but without searching all rules. - Improve parser, so that it warns about bad rules. These are: - "tablearg" argument to action, but no "table" in the rule. All traffic will be blocked. - "tablearg" argument to action, but "table" searches for entry with a specific value. All traffic will be blocked. - "tablearg" argument to action, and two "table" looks - for src and for dst. The last lookup will match.
2005-12-13 12:16:03 +00:00
} else
errx(EX_DATAERR, "illegal argument for %s", *(av - 1));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add a ng_ipfw node, implementing a quick and simple interface between ipfw(4) and netgraph(4) facilities. Reviewed by: andre, brooks, julian
2005-02-05 12:06:33 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_FORWARD: {
ipfw_insn_sa *p = (ipfw_insn_sa *)action;
char *s, *end;
NEED1("missing forward address[:port]");
action->opcode = O_FORWARD_IP;
action->len = F_INSN_SIZE(ipfw_insn_sa);
Explain that we assume AF_INET and only use the addr and port field from a struct sockaddr_in, so there is no need to initialize sin_len
2009-02-02 11:02:19 +00:00
/*
* In the kernel we assume AF_INET and use only
restore setting of sin_len (was removed in 1.146 last february) as it seems that now it is necessary for 'forward' to work outside lo0. The bug (and fix) was reported on 8.0. This patch probably applies to RELENG_7 as well. It seems that 'pf' has a similar bug. Submitted by: Lytochkin Boris MFC after: 3 days
2009-12-06 18:04:26 +00:00
* sin_port and sin_addr. Remember to set sin_len as
* the routing code seems to use it too.
Explain that we assume AF_INET and only use the addr and port field from a struct sockaddr_in, so there is no need to initialize sin_len
2009-02-02 11:02:19 +00:00
*/
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
p->sa.sin_family = AF_INET;
restore setting of sin_len (was removed in 1.146 last february) as it seems that now it is necessary for 'forward' to work outside lo0. The bug (and fix) was reported on 8.0. This patch probably applies to RELENG_7 as well. It seems that 'pf' has a similar bug. Submitted by: Lytochkin Boris MFC after: 3 days
2009-12-06 18:04:26 +00:00
p->sa.sin_len = sizeof(struct sockaddr_in);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
p->sa.sin_port = 0;
/*
* locate the address-port separator (':' or ',')
*/
s = strchr(*av, ':');
if (s == NULL)
s = strchr(*av, ',');
if (s != NULL) {
*(s++) = '\0';
i = strtoport(s, &end, 0 /* base */, 0 /* proto */);
if (s == end)
errx(EX_DATAERR,
"illegal forwarding port ``%s''", s);
Store the port number in "fwd" rules in host format, same as ipfw1 has always done. Technically, this is the wrong format, but it reduces the diffs in -stable. Someday, when we get rid of ipfw1, I will put the port number in the proper format both in kernel and userland. MFC after: 3 days (with re@ permission)
2002-09-12 00:45:32 +00:00
p->sa.sin_port = (u_short)i;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
comply with style police Submitted by: ru MFC after: 1 month
2006-08-18 22:36:05 +00:00
if (_substrcmp(*av, "tablearg") == 0)
p->sa.sin_addr.s_addr = INADDR_ANY;
else
Allow ipfw to forward to a destination that is specified by a table. for example: fwd tablearg ip from any to table(1) where table 1 has entries of the form: 1.1.1.0/24 10.2.3.4 208.23.2.0/24 router2 This allows trivial implementation of a secondary routing table implemented in the firewall layer. I expect more work (under discussion with Glebius) to follow this to clean up some of the messy parts of ipfw related to tables. Reviewed by: Glebius MFC after: 1 month
2006-08-17 22:49:50 +00:00
lookup_host(*av, &(p->sa.sin_addr));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
Allow ipfw to forward to a destination that is specified by a table. for example: fwd tablearg ip from any to table(1) where table 1 has entries of the form: 1.1.1.0/24 10.2.3.4 208.23.2.0/24 router2 This allows trivial implementation of a secondary routing table implemented in the firewall layer. I expect more work (under discussion with Glebius) to follow this to clean up some of the messy parts of ipfw related to tables. Reviewed by: Glebius MFC after: 1 month
2006-08-17 22:49:50 +00:00
}
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
case TOK_COMMENT:
/* pretend it is a 'count' rule followed by the comment */
action->opcode = O_COUNT;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av--; /* go back... */
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
break;
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
case TOK_SETFIB:
{
int numfibs;
Change two variables to size_t to improve portability. Submitted by: Xin Li
2008-05-10 15:02:56 +00:00
size_t intsize = sizeof(int);
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
action->opcode = O_SETFIB;
NEED1("missing fib number");
action->arg1 = strtoul(*av, NULL, 10);
Change two variables to size_t to improve portability. Submitted by: Xin Li
2008-05-10 15:02:56 +00:00
if (sysctlbyname("net.fibs", &numfibs, &intsize, NULL, 0) == -1)
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
errx(EX_DATAERR, "fibs not suported.\n");
if (action->arg1 >= numfibs) /* Temporary */
errx(EX_DATAERR, "fib too large.\n");
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
break;
}
Implement an ipfw action to reassemble ip packets: reass.
2009-04-01 20:23:47 +00:00
case TOK_REASS:
action->opcode = O_REASS;
break;
Summer of Code 2005: improve libalias - part 2 of 2 With the second (and last) part of my previous Summer of Code work, we get: -ipfw's in kernel nat -redirect_* and LSNAT support General information about nat syntax and some examples are available in the ipfw (8) man page. The redirect and LSNAT syntax are identical to natd, so please refer to natd (8) man page. To enable in kernel nat in rc.conf, two options were added: o firewall_nat_enable: equivalent to natd_enable o firewall_nat_interface: equivalent to natd_interface Remember to set net.inet.ip.fw.one_pass to 0, if you want the packet to continue being checked by the firewall ruleset after being (de)aliased. NOTA BENE: due to some problems with libalias architecture, in kernel nat won't work with TSO enabled nic, thus you have to disable TSO via ifconfig (ifconfig foo0 -tso). Approved by: glebius (mentor)
2006-12-29 21:59:17 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
default:
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
errx(EX_DATAERR, "invalid action %s\n", av[-1]);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
action = next_cmd(action);
/*
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
* [altq queuename] -- altq tag, optional
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* [log [logamount N]] -- log, optional
*
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
* If they exist, it go first in the cmdbuf, but then it is
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* skipped in the copy section to the end of the buffer.
*/
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
while (av[0] != NULL && (i = match_token(rule_action_params, *av)) != -1) {
av++;
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
switch (i) {
case TOK_LOG:
{
ipfw_insn_log *c = (ipfw_insn_log *)cmd;
int l;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
if (have_log)
errx(EX_DATAERR,
"log cannot be specified more than once");
have_log = (ipfw_insn *)c;
cmd->len = F_INSN_SIZE(ipfw_insn_log);
cmd->opcode = O_LOG;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && _substrcmp(*av, "logamount") == 0) {
av++;
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
NEED1("logamount requires argument");
l = atoi(*av);
if (l < 0)
errx(EX_DATAERR,
"logamount must be positive");
c->max_log = l;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
} else {
len = sizeof(c->max_log);
if (sysctlbyname("net.inet.ip.fw.verbose_limit",
&c->max_log, &len, NULL, 0) == -1)
errx(1, "sysctlbyname(\"%s\")",
"net.inet.ip.fw.verbose_limit");
}
}
break;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#ifndef NO_ALTQ
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
case TOK_ALTQ:
{
ipfw_insn_altq *a = (ipfw_insn_altq *)cmd;
NEED1("missing altq queue name");
if (have_altq)
errx(EX_DATAERR,
"altq cannot be specified more than once");
have_altq = (ipfw_insn *)a;
cmd->len = F_INSN_SIZE(ipfw_insn_altq);
cmd->opcode = O_ALTQ;
put the altq-related functions into a separate file. Minor cleanup of the includes used by the various source files, including annotations of why certain headers are used.
2009-02-01 16:00:49 +00:00
a->qid = altq_name_to_qid(*av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
}
break;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#endif
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
case TOK_TAG:
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
case TOK_UNTAG: {
uint16_t tag;
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
if (have_tag)
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
errx(EX_USAGE, "tag and untag cannot be "
"specified more than once");
Several ipfw options and actions use a 16-bit argument to indicate pipes, queues, tags, rule numbers and so on. These are all different namespaces, and the only thing they have in common is the fact they use a 16-bit slot to represent the argument. There is some confusion in the code, mostly for historical reasons, on how the values 0 and 65535 should be used. At the moment, 0 is forbidden almost everywhere, while 65535 is used to represent a 'tablearg' argument, i.e. the result of the most recent table() lookup. For now, try to use explicit constants for the min and max allowed values, and do not overload the default rule number for that. Also, make the MTAG_IPFW declaration only visible to the kernel. NOTE: I think the issue needs to be revisited before 8.0 is out: the 2^16 namespace limit for rule numbers and pipe/queue is annoying, and we can easily bump the limit to 2^32 which gives a lot more flexibility in partitioning the namespace. MFC after: 5 days
2009-06-05 16:16:07 +00:00
GET_UINT_ARG(tag, IPFW_ARG_MIN, IPFW_ARG_MAX, i,
Use IPFW_DEFAULT_RULE instead of hardcoded value since now it is available. MFC after: 5 days.
2008-09-06 17:23:37 +00:00
rule_action_params);
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
have_tag = cmd;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
fill_cmd(cmd, O_TAG, (i == TOK_TAG) ? 0: F_NOT, tag);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
break;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
}
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
default:
abort();
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
cmd = next_cmd(cmd);
}
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
if (have_state) /* must be a check-state, we are done */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
goto done;
#define OR_START(target) \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && (*av[0] == '(' || *av[0] == '{')) { \
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (open_par) \
errx(EX_USAGE, "nested \"(\" not allowed\n"); \
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
prev = NULL; \
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
open_par = 1; \
if ( (av[0])[1] == '\0') { \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++; \
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
} else \
(*av)++; \
} \
target: \
#define CLOSE_PAR \
if (open_par) { \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && ( \
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
strcmp(*av, ")") == 0 || \
strcmp(*av, "}") == 0)) { \
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
prev = NULL; \
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
open_par = 0; \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++; \
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
} else \
errx(EX_USAGE, "missing \")\"\n"); \
}
Kill EOL whitespaces, style(9) fix.
2002-11-06 15:09:34 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#define NOT_BLOCK \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && _substrcmp(*av, "not") == 0) { \
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (cmd->len & F_NOT) \
errx(EX_USAGE, "double \"not\" not allowed\n"); \
cmd->len |= F_NOT; \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++; \
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
#define OR_BLOCK(target) \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && _substrcmp(*av, "or") == 0) { \
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (prev == NULL || open_par == 0) \
errx(EX_DATAERR, "invalid OR block"); \
prev->len |= F_OR; \
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++; \
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
goto target; \
} \
CLOSE_PAR;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
first_cmd = cmd;
One more (hopefully the last one) step in cleaning up the syntax, following Julian's good suggestion: since you can specify any match pattern as an option, rules now have the following format: [<proto> from <src> to <dst>] [options] i.e. the first part is now entirely optional (and left there just for compatibility with ipfw1 rulesets). Add a "-c" flag to show/list rules in the compact form (i.e. without the "ip from any to any" part) when possible. The default is to include it so that scripts processing ipfw's canonical output will still work. Note that as part of this cleanup (and to remove ambiguity), MAC fields now can only be specified in the options part. Update the manpage to reflect the syntax. Clarify the behaviour when a match is attempted on fields which are not present in the packet, e.g. port numbers on non TCP/UDP packets, and the "not" operator is specified. E.g. ipfw add allow not src-port 80 will match also ICMP packets because they do not have port numbers, so "src-port 80" will fail and "not src-port 80" will succeed. For such cases it is advised to insert further options to prevent undesired results (e.g. in the case above, "ipfw add allow proto tcp not src-port 80"). We definitely need to rewrite the parser using lex and yacc!
2002-08-19 12:36:54 +00:00
#if 0
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
/*
* MAC addresses, optional.
* If we have this, we skip the part "proto from src to dst"
* and jump straight to the option parsing.
*/
NOT_BLOCK;
NEED1("missing protocol");
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(*av, "MAC") == 0 ||
_substrcmp(*av, "mac") == 0) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++; /* the "MAC" keyword */
add_mac(cmd, av); /* exits in case of errors */
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
cmd = next_cmd(cmd);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av += 2; /* dst-mac and src-mac */
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
NOT_BLOCK;
NEED1("missing mac type");
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (add_mactype(cmd, av[0]))
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
cmd = next_cmd(cmd);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++; /* any or mac-type */
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
goto read_options;
}
One more (hopefully the last one) step in cleaning up the syntax, following Julian's good suggestion: since you can specify any match pattern as an option, rules now have the following format: [<proto> from <src> to <dst>] [options] i.e. the first part is now entirely optional (and left there just for compatibility with ipfw1 rulesets). Add a "-c" flag to show/list rules in the compact form (i.e. without the "ip from any to any" part) when possible. The default is to include it so that scripts processing ipfw's canonical output will still work. Note that as part of this cleanup (and to remove ambiguity), MAC fields now can only be specified in the options part. Update the manpage to reflect the syntax. Clarify the behaviour when a match is attempted on fields which are not present in the packet, e.g. port numbers on non TCP/UDP packets, and the "not" operator is specified. E.g. ipfw add allow not src-port 80 will match also ICMP packets because they do not have port numbers, so "src-port 80" will fail and "not src-port 80" will succeed. For such cases it is advised to insert further options to prevent undesired results (e.g. in the case above, "ipfw add allow proto tcp not src-port 80"). We definitely need to rewrite the parser using lex and yacc!
2002-08-19 12:36:54 +00:00
#endif
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* protocol, mandatory
*/
OR_START(get_proto);
NOT_BLOCK;
NEED1("missing protocol");
We couldn't specify the rule for filtering tunnel traffic since an IPv6 support was committed: - Stop treating `ip' and `ipv6' as special in `proto' option as they conflict with /etc/protocols. - Disuse `ipv4' in `proto' option as it is corresponding to `ipv6'. - When protocol is specified as numeric, treat it as it is even it is 41 (ipv6). - Allow zero for protocol as it is valid number of `ip'. Still, we cannot specify an IPv6 over an IPv4 tunnel like before such as: pass ipv6 from any to any But, now, you can specify it like: pass ip4 from any to any proto ipv6 PR: kern/89472 Reported by: Ga l Roualland <gael.roualland__at__dial.oleane.com> MFC after: 1 week
2005-11-29 15:25:09 +00:00
if (add_proto_compat(cmd, *av, &proto)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
add_proto() now fills proto for us so stop to 'guess' the protocol from the command and rather trust the value add_proto filled in. While here, fix an oversight in the pretty printing of ip6/4 options.
2005-06-07 14:11:17 +00:00
if (F_LEN(cmd) != 0) {
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
prev = cmd;
cmd = next_cmd(cmd);
}
One more (hopefully the last one) step in cleaning up the syntax, following Julian's good suggestion: since you can specify any match pattern as an option, rules now have the following format: [<proto> from <src> to <dst>] [options] i.e. the first part is now entirely optional (and left there just for compatibility with ipfw1 rulesets). Add a "-c" flag to show/list rules in the compact form (i.e. without the "ip from any to any" part) when possible. The default is to include it so that scripts processing ipfw's canonical output will still work. Note that as part of this cleanup (and to remove ambiguity), MAC fields now can only be specified in the options part. Update the manpage to reflect the syntax. Clarify the behaviour when a match is attempted on fields which are not present in the packet, e.g. port numbers on non TCP/UDP packets, and the "not" operator is specified. E.g. ipfw add allow not src-port 80 will match also ICMP packets because they do not have port numbers, so "src-port 80" will fail and "not src-port 80" will succeed. For such cases it is advised to insert further options to prevent undesired results (e.g. in the case above, "ipfw add allow proto tcp not src-port 80"). We definitely need to rewrite the parser using lex and yacc!
2002-08-19 12:36:54 +00:00
} else if (first_cmd != cmd) {
o Pass a correct argument to printf(3). PR: bin/51750 Submitted by: Vasil Dimov <vd@datamax.bg> MFC after: 2 weeks
2003-06-16 09:44:53 +00:00
errx(EX_DATAERR, "invalid protocol ``%s''", *av);
One more (hopefully the last one) step in cleaning up the syntax, following Julian's good suggestion: since you can specify any match pattern as an option, rules now have the following format: [<proto> from <src> to <dst>] [options] i.e. the first part is now entirely optional (and left there just for compatibility with ipfw1 rulesets). Add a "-c" flag to show/list rules in the compact form (i.e. without the "ip from any to any" part) when possible. The default is to include it so that scripts processing ipfw's canonical output will still work. Note that as part of this cleanup (and to remove ambiguity), MAC fields now can only be specified in the options part. Update the manpage to reflect the syntax. Clarify the behaviour when a match is attempted on fields which are not present in the packet, e.g. port numbers on non TCP/UDP packets, and the "not" operator is specified. E.g. ipfw add allow not src-port 80 will match also ICMP packets because they do not have port numbers, so "src-port 80" will fail and "not src-port 80" will succeed. For such cases it is advised to insert further options to prevent undesired results (e.g. in the case above, "ipfw add allow proto tcp not src-port 80"). We definitely need to rewrite the parser using lex and yacc!
2002-08-19 12:36:54 +00:00
} else
goto read_options;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
OR_BLOCK(get_proto);
/*
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
* "from", mandatory
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if ((av[0] == NULL) || _substrcmp(*av, "from") != 0)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
errx(EX_USAGE, "missing ``from''");
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* source IP, mandatory
*/
OR_START(source_ip);
NOT_BLOCK; /* optional "not" */
NEED1("missing source address");
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
if (add_src(cmd, *av, proto)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
if (F_LEN(cmd) != 0) { /* ! any */
prev = cmd;
cmd = next_cmd(cmd);
}
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
} else
errx(EX_USAGE, "bad source address %s", *av);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
OR_BLOCK(source_ip);
/*
* source ports, optional
*/
NOT_BLOCK; /* optional "not" */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if ( av[0] != NULL ) {
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(*av, "any") == 0 ||
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
add_ports(cmd, *av, proto, O_IP_SRCPORT)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
if (F_LEN(cmd) != 0)
cmd = next_cmd(cmd);
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
/*
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
* "to", mandatory
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if ( (av[0] == NULL) || _substrcmp(*av, "to") != 0 )
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
errx(EX_USAGE, "missing ``to''");
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* destination, mandatory
*/
OR_START(dest_ip);
NOT_BLOCK; /* optional "not" */
NEED1("missing dst address");
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
if (add_dst(cmd, *av, proto)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
if (F_LEN(cmd) != 0) { /* ! any */
prev = cmd;
cmd = next_cmd(cmd);
}
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
} else
errx( EX_USAGE, "bad destination address %s", *av);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
OR_BLOCK(dest_ip);
/*
* dest. ports, optional
*/
NOT_BLOCK; /* optional "not" */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0]) {
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(*av, "any") == 0 ||
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
add_ports(cmd, *av, proto, O_IP_DSTPORT)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
if (F_LEN(cmd) != 0)
cmd = next_cmd(cmd);
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
read_options:
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && first_cmd == cmd) {
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
/*
* nothing specified so far, store in the rule to ease
* printout later.
*/
rule->_pad = 1;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
prev = NULL;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
while ( av[0] != NULL ) {
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
char *s;
ipfw_insn_u32 *cmd32; /* alias for cmd */
s = *av;
cmd32 = (ipfw_insn_u32 *)cmd;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (*s == '!') { /* alternate syntax for NOT */
if (cmd->len & F_NOT)
errx(EX_USAGE, "double \"not\" not allowed\n");
cmd->len = F_NOT;
s++;
}
i = match_token(rule_options, s);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
switch(i) {
case TOK_NOT:
if (cmd->len & F_NOT)
errx(EX_USAGE, "double \"not\" not allowed\n");
cmd->len = F_NOT;
break;
case TOK_OR:
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
if (open_par == 0 || prev == NULL)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
errx(EX_USAGE, "invalid \"or\" block\n");
prev->len |= F_OR;
break;
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
case TOK_STARTBRACE:
if (open_par)
errx(EX_USAGE, "+nested \"(\" not allowed\n");
open_par = 1;
break;
case TOK_ENDBRACE:
if (!open_par)
errx(EX_USAGE, "+missing \")\"\n");
open_par = 0;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
prev = NULL;
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_IN:
fill_cmd(cmd, O_IN, 0, 0);
break;
case TOK_OUT:
cmd->len ^= F_NOT; /* toggle F_NOT */
fill_cmd(cmd, O_IN, 0, 0);
break;
Add support to IPFW for classification based on "diverted" status (that is, input via a divert socket).
2004-10-03 00:26:35 +00:00
case TOK_DIVERTED:
fill_cmd(cmd, O_DIVERTED, 0, 3);
break;
case TOK_DIVERTEDLOOPBACK:
fill_cmd(cmd, O_DIVERTED, 0, 1);
break;
case TOK_DIVERTEDOUTPUT:
fill_cmd(cmd, O_DIVERTED, 0, 2);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_FRAG:
fill_cmd(cmd, O_FRAG, 0, 0);
break;
case TOK_LAYER2:
fill_cmd(cmd, O_LAYER2, 0, 0);
break;
case TOK_XMIT:
case TOK_RECV:
case TOK_VIA:
NEED1("recv, xmit, via require interface name"
" or address");
fill_iface((ipfw_insn_if *)cmd, av[0]);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (F_LEN(cmd) == 0) /* not a valid address */
break;
if (i == TOK_XMIT)
cmd->opcode = O_XMIT;
else if (i == TOK_RECV)
cmd->opcode = O_RECV;
else if (i == TOK_VIA)
cmd->opcode = O_VIA;
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case TOK_ICMPTYPES:
NEED1("icmptypes requires list of types");
fill_icmptypes((ipfw_insn_u32 *)cmd, *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
break;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
case TOK_ICMP6TYPES:
NEED1("icmptypes requires list of types");
fill_icmp6types((ipfw_insn_icmp6 *)cmd, *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_IPTTL:
NEED1("ipttl requires TTL");
Add support for multiple values and ranges for the "iplen", "ipttl", "ipid" options. This feature has been requested by several users. On passing, fix some minor bugs in the parser. This change is fully backward compatible so if you have an old /sbin/ipfw and a new kernel you are not in trouble (but you need to update /sbin/ipfw if you want to use the new features). Document the changes in the manpage. Now you can write things like ipfw add skipto 1000 iplen 0-500 which some people were asking to give preferential treatment to short packets. The 'MFC after' is just set as a reminder, because I still need to merge the Alpha/Sparc64 fixes for ipfw2 (which unfortunately change the size of certain kernel structures; not that it matters a lot since ipfw2 is entirely optional and not the default...) PR: bin/48015 MFC after: 1 week
2003-06-22 17:33:19 +00:00
if (strpbrk(*av, "-,")) {
if (!add_ports(cmd, *av, 0, O_IPTTL))
errx(EX_DATAERR, "invalid ipttl %s", *av);
} else
fill_cmd(cmd, O_IPTTL, 0, strtoul(*av, NULL, 0));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case TOK_IPID:
Add support for multiple values and ranges for the "iplen", "ipttl", "ipid" options. This feature has been requested by several users. On passing, fix some minor bugs in the parser. This change is fully backward compatible so if you have an old /sbin/ipfw and a new kernel you are not in trouble (but you need to update /sbin/ipfw if you want to use the new features). Document the changes in the manpage. Now you can write things like ipfw add skipto 1000 iplen 0-500 which some people were asking to give preferential treatment to short packets. The 'MFC after' is just set as a reminder, because I still need to merge the Alpha/Sparc64 fixes for ipfw2 (which unfortunately change the size of certain kernel structures; not that it matters a lot since ipfw2 is entirely optional and not the default...) PR: bin/48015 MFC after: 1 week
2003-06-22 17:33:19 +00:00
NEED1("ipid requires id");
if (strpbrk(*av, "-,")) {
if (!add_ports(cmd, *av, 0, O_IPID))
errx(EX_DATAERR, "invalid ipid %s", *av);
} else
fill_cmd(cmd, O_IPID, 0, strtoul(*av, NULL, 0));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case TOK_IPLEN:
NEED1("iplen requires length");
Add support for multiple values and ranges for the "iplen", "ipttl", "ipid" options. This feature has been requested by several users. On passing, fix some minor bugs in the parser. This change is fully backward compatible so if you have an old /sbin/ipfw and a new kernel you are not in trouble (but you need to update /sbin/ipfw if you want to use the new features). Document the changes in the manpage. Now you can write things like ipfw add skipto 1000 iplen 0-500 which some people were asking to give preferential treatment to short packets. The 'MFC after' is just set as a reminder, because I still need to merge the Alpha/Sparc64 fixes for ipfw2 (which unfortunately change the size of certain kernel structures; not that it matters a lot since ipfw2 is entirely optional and not the default...) PR: bin/48015 MFC after: 1 week
2003-06-22 17:33:19 +00:00
if (strpbrk(*av, "-,")) {
if (!add_ports(cmd, *av, 0, O_IPLEN))
errx(EX_DATAERR, "invalid ip len %s", *av);
} else
fill_cmd(cmd, O_IPLEN, 0, strtoul(*av, NULL, 0));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case TOK_IPVER:
NEED1("ipver requires version");
fill_cmd(cmd, O_IPVER, 0, strtoul(*av, NULL, 0));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case TOK_IPPRECEDENCE:
NEED1("ipprecedence requires value");
fill_cmd(cmd, O_IPPRECEDENCE, 0,
(strtoul(*av, NULL, 0) & 7) << 5);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_IPOPTS:
NEED1("missing argument for ipoptions");
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
fill_flags(cmd, O_IPOPT, f_ipopts, *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case TOK_IPTOS:
NEED1("missing argument for iptos");
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
fill_flags(cmd, O_IPTOS, f_iptos, *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_UID:
NEED1("uid requires argument");
{
char *end;
uid_t uid;
struct passwd *pwd;
cmd->opcode = O_UID;
uid = strtoul(*av, &end, 0);
pwd = (*end == '\0') ? getpwuid(uid) : getpwnam(*av);
if (pwd == NULL)
errx(EX_DATAERR, "uid \"%s\" nonexistent", *av);
Fix UID/GID options parsing. PR: bin/42579 Submitted by: Belousov Oleg <oleg@belousov.com> Approved by: luigi MFC after: 2 weeks
2002-11-06 14:57:18 +00:00
cmd32->d[0] = pwd->pw_uid;
Currently when ipfw(8) generates the micro-instructions for rules which contain O_UID, O_GID and O_JAIL opcodes, the F_NOT or F_OR logical operator bits get clobbered. Making it impossible to use the ``NOT'' or ``OR'' operators with uid, gid and jail based constraints. The ipfw_insn instruction template contains a ``len'' element which stores two pieces of information, the size of the instruction (in 32-bit words) in the low 6 bits of "len" with the 2 remaining bits to implement OR and NOT. The current code clobbers the OR and NOT bits by initializing the ``len'' element to the size, rather than OR'ing the bits. This change fixes this by changing the initialization of cmd->len to an OR operation for the O_UID, O_GID and O_JAIL opcodes. This may be a MFC candidate for RELENG_5. Reviewed by: andre Approved by: luigi PR: kern/63961 (partially)
2004-09-11 19:44:29 +00:00
cmd->len |= F_INSN_SIZE(ipfw_insn_u32);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
break;
case TOK_GID:
NEED1("gid requires argument");
{
char *end;
gid_t gid;
struct group *grp;
cmd->opcode = O_GID;
gid = strtoul(*av, &end, 0);
grp = (*end == '\0') ? getgrgid(gid) : getgrnam(*av);
if (grp == NULL)
errx(EX_DATAERR, "gid \"%s\" nonexistent", *av);
Fix UID/GID options parsing. PR: bin/42579 Submitted by: Belousov Oleg <oleg@belousov.com> Approved by: luigi MFC after: 2 weeks
2002-11-06 14:57:18 +00:00
cmd32->d[0] = grp->gr_gid;
Currently when ipfw(8) generates the micro-instructions for rules which contain O_UID, O_GID and O_JAIL opcodes, the F_NOT or F_OR logical operator bits get clobbered. Making it impossible to use the ``NOT'' or ``OR'' operators with uid, gid and jail based constraints. The ipfw_insn instruction template contains a ``len'' element which stores two pieces of information, the size of the instruction (in 32-bit words) in the low 6 bits of "len" with the 2 remaining bits to implement OR and NOT. The current code clobbers the OR and NOT bits by initializing the ``len'' element to the size, rather than OR'ing the bits. This change fixes this by changing the initialization of cmd->len to an OR operation for the O_UID, O_GID and O_JAIL opcodes. This may be a MFC candidate for RELENG_5. Reviewed by: andre Approved by: luigi PR: kern/63961 (partially)
2004-09-11 19:44:29 +00:00
cmd->len |= F_INSN_SIZE(ipfw_insn_u32);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
break;
Add the ability to associate ipfw rules with a specific prison ID. Since the only thing truly unique about a prison is it's ID, I figured this would be the most granular way of handling this. This commit makes the following changes: - Adds tokenizing and parsing for the ``jail'' command line option to the ipfw(8) userspace utility. - Append the ipfw opcode list with O_JAIL. - While Iam here, add a comment informing others that if they want to add additional opcodes, they should append them to the end of the list to avoid ABI breakage. - Add ``fw_prid'' to the ipfw ucred cache structure. - When initializing ucred cache, if the process is jailed, set fw_prid to the prison ID, otherwise set it to -1. - Update man page to reflect these changes. This change was a strong motivator behind the ucred caching mechanism in ipfw. A sample usage of this new functionality could be: ipfw add count ip from any to any jail 2 It should be noted that because ucred based constraints are only implemented for TCP and UDP packets, the same applies for jail associations. Conceptual head nod by: pjd Reviewed by: rwatson Approved by: bmilekic (mentor)
2004-08-12 22:06:55 +00:00
case TOK_JAIL:
NEED1("jail requires argument");
{
char *end;
int jid;
cmd->opcode = O_JAIL;
jid = (int)strtol(*av, &end, 0);
if (jid < 0 || *end != '\0')
errx(EX_DATAERR, "jail requires prison ID");
Since "d" is an array of 32 bit values, it is more correct to change the cast from unsigned int to uint32_t. Pointed out by: luigi
2004-09-21 22:12:43 +00:00
cmd32->d[0] = (uint32_t)jid;
Currently when ipfw(8) generates the micro-instructions for rules which contain O_UID, O_GID and O_JAIL opcodes, the F_NOT or F_OR logical operator bits get clobbered. Making it impossible to use the ``NOT'' or ``OR'' operators with uid, gid and jail based constraints. The ipfw_insn instruction template contains a ``len'' element which stores two pieces of information, the size of the instruction (in 32-bit words) in the low 6 bits of "len" with the 2 remaining bits to implement OR and NOT. The current code clobbers the OR and NOT bits by initializing the ``len'' element to the size, rather than OR'ing the bits. This change fixes this by changing the initialization of cmd->len to an OR operation for the O_UID, O_GID and O_JAIL opcodes. This may be a MFC candidate for RELENG_5. Reviewed by: andre Approved by: luigi PR: kern/63961 (partially)
2004-09-11 19:44:29 +00:00
cmd->len |= F_INSN_SIZE(ipfw_insn_u32);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add the ability to associate ipfw rules with a specific prison ID. Since the only thing truly unique about a prison is it's ID, I figured this would be the most granular way of handling this. This commit makes the following changes: - Adds tokenizing and parsing for the ``jail'' command line option to the ipfw(8) userspace utility. - Append the ipfw opcode list with O_JAIL. - While Iam here, add a comment informing others that if they want to add additional opcodes, they should append them to the end of the list to avoid ABI breakage. - Add ``fw_prid'' to the ipfw ucred cache structure. - When initializing ucred cache, if the process is jailed, set fw_prid to the prison ID, otherwise set it to -1. - Update man page to reflect these changes. This change was a strong motivator behind the ucred caching mechanism in ipfw. A sample usage of this new functionality could be: ipfw add count ip from any to any jail 2 It should be noted that because ucred based constraints are only implemented for TCP and UDP packets, the same applies for jail associations. Conceptual head nod by: pjd Reviewed by: rwatson Approved by: bmilekic (mentor)
2004-08-12 22:06:55 +00:00
}
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_ESTAB:
fill_cmd(cmd, O_ESTAB, 0, 0);
break;
case TOK_SETUP:
fill_cmd(cmd, O_TCPFLAGS, 0,
(TH_SYN) | ( (TH_ACK) & 0xff) <<8 );
break;
Add support to IPFW for matching by TCP data length.
2004-10-03 00:47:15 +00:00
case TOK_TCPDATALEN:
NEED1("tcpdatalen requires length");
if (strpbrk(*av, "-,")) {
if (!add_ports(cmd, *av, 0, O_TCPDATALEN))
errx(EX_DATAERR, "invalid tcpdata len %s", *av);
} else
fill_cmd(cmd, O_TCPDATALEN, 0,
strtoul(*av, NULL, 0));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add support to IPFW for matching by TCP data length.
2004-10-03 00:47:15 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case TOK_TCPOPTS:
NEED1("missing argument for tcpoptions");
fill_flags(cmd, O_TCPOPTS, f_tcpopts, *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case TOK_TCPSEQ:
case TOK_TCPACK:
NEED1("tcpseq/tcpack requires argument");
cmd->len = F_INSN_SIZE(ipfw_insn_u32);
cmd->opcode = (i == TOK_TCPSEQ) ? O_TCPSEQ : O_TCPACK;
cmd32->d[0] = htonl(strtoul(*av, NULL, 0));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case TOK_TCPWIN:
NEED1("tcpwin requires length");
fill_cmd(cmd, O_TCPWIN, 0,
htons(strtoul(*av, NULL, 0)));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case TOK_TCPFLAGS:
NEED1("missing argument for tcpflags");
cmd->opcode = O_TCPFLAGS;
fill_flags(cmd, O_TCPFLAGS, f_tcpflags, *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case TOK_KEEPSTATE:
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
if (open_par)
errx(EX_USAGE, "keep-state cannot be part "
"of an or block");
A bunch of minor fixes: * accept "icmptype" as an alias for "icmptypes"; * remove an extra whitespace after "log" rules; * print correctly the "limit" masks; * correct a typo in parsing dummynet arguments (this caused a coredump); * do not allow specifying both "check-state" and "limit", they are (and have always been) mutually exclusive; * remove an extra print of the rule before installing it; * make stdout buffered -- otherwise, if you log its output with syslog, you will see one entry for each printf(). Rather unpleasant.
2002-07-13 15:57:23 +00:00
if (have_state)
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
errx(EX_USAGE, "only one of keep-state "
A bunch of minor fixes: * accept "icmptype" as an alias for "icmptypes"; * remove an extra whitespace after "log" rules; * print correctly the "limit" masks; * correct a typo in parsing dummynet arguments (this caused a coredump); * do not allow specifying both "check-state" and "limit", they are (and have always been) mutually exclusive; * remove an extra print of the rule before installing it; * make stdout buffered -- otherwise, if you log its output with syslog, you will see one entry for each printf(). Rather unpleasant.
2002-07-13 15:57:23 +00:00
"and limit is allowed");
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
have_state = cmd;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
fill_cmd(cmd, O_KEEP_STATE, 0, 0);
break;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
case TOK_LIMIT: {
ipfw_insn_limit *c = (ipfw_insn_limit *)cmd;
int val;
Fix one parsing bug introduced by last commit, and correct parsing and printing of or-blocks in address, ports and options lists.
2002-08-10 15:10:15 +00:00
if (open_par)
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
errx(EX_USAGE,
"limit cannot be part of an or block");
A bunch of minor fixes: * accept "icmptype" as an alias for "icmptypes"; * remove an extra whitespace after "log" rules; * print correctly the "limit" masks; * correct a typo in parsing dummynet arguments (this caused a coredump); * do not allow specifying both "check-state" and "limit", they are (and have always been) mutually exclusive; * remove an extra print of the rule before installing it; * make stdout buffered -- otherwise, if you log its output with syslog, you will see one entry for each printf(). Rather unpleasant.
2002-07-13 15:57:23 +00:00
if (have_state)
o Add missed w/space in the error message. Spotted by: Ivan Voras MFC after: 1 week
2007-04-17 16:36:24 +00:00
errx(EX_USAGE, "only one of keep-state and "
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
"limit is allowed");
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
have_state = cmd;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->len = F_INSN_SIZE(ipfw_insn_limit);
cmd->opcode = O_LIMIT;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
c->limit_mask = c->conn_limit = 0;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
while ( av[0] != NULL ) {
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
if ((val = match_token(limit_masks, *av)) <= 0)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
c->limit_mask |= val;
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (c->limit_mask == 0)
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
errx(EX_USAGE, "limit: missing limit mask");
Several ipfw options and actions use a 16-bit argument to indicate pipes, queues, tags, rule numbers and so on. These are all different namespaces, and the only thing they have in common is the fact they use a 16-bit slot to represent the argument. There is some confusion in the code, mostly for historical reasons, on how the values 0 and 65535 should be used. At the moment, 0 is forbidden almost everywhere, while 65535 is used to represent a 'tablearg' argument, i.e. the result of the most recent table() lookup. For now, try to use explicit constants for the min and max allowed values, and do not overload the default rule number for that. Also, make the MTAG_IPFW declaration only visible to the kernel. NOTE: I think the issue needs to be revisited before 8.0 is out: the 2^16 namespace limit for rule numbers and pipe/queue is annoying, and we can easily bump the limit to 2^32 which gives a lot more flexibility in partitioning the namespace. MFC after: 5 days
2009-06-05 16:16:07 +00:00
GET_UINT_ARG(c->conn_limit, IPFW_ARG_MIN, IPFW_ARG_MAX,
Use IPFW_DEFAULT_RULE instead of hardcoded value since now it is available. MFC after: 5 days.
2008-09-06 17:23:37 +00:00
TOK_LIMIT, rule_options);
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
case TOK_PROTO:
NEED1("missing protocol");
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
if (add_proto(cmd, *av, &proto)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
One more (hopefully the last one) step in cleaning up the syntax, following Julian's good suggestion: since you can specify any match pattern as an option, rules now have the following format: [<proto> from <src> to <dst>] [options] i.e. the first part is now entirely optional (and left there just for compatibility with ipfw1 rulesets). Add a "-c" flag to show/list rules in the compact form (i.e. without the "ip from any to any" part) when possible. The default is to include it so that scripts processing ipfw's canonical output will still work. Note that as part of this cleanup (and to remove ambiguity), MAC fields now can only be specified in the options part. Update the manpage to reflect the syntax. Clarify the behaviour when a match is attempted on fields which are not present in the packet, e.g. port numbers on non TCP/UDP packets, and the "not" operator is specified. E.g. ipfw add allow not src-port 80 will match also ICMP packets because they do not have port numbers, so "src-port 80" will fail and "not src-port 80" will succeed. For such cases it is advised to insert further options to prevent undesired results (e.g. in the case above, "ipfw add allow proto tcp not src-port 80"). We definitely need to rewrite the parser using lex and yacc!
2002-08-19 12:36:54 +00:00
} else
o Pass a correct argument to printf(3). PR: bin/51750 Submitted by: Vasil Dimov <vd@datamax.bg> MFC after: 2 weeks
2003-06-16 09:44:53 +00:00
errx(EX_DATAERR, "invalid protocol ``%s''",
*av);
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
break;
Kill EOL whitespaces, style(9) fix.
2002-11-06 15:09:34 +00:00
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
case TOK_SRCIP:
NEED1("missing source IP");
if (add_srcip(cmd, *av)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
}
break;
case TOK_DSTIP:
NEED1("missing destination IP");
if (add_dstip(cmd, *av)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
}
break;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
case TOK_SRCIP6:
NEED1("missing source IP6");
if (add_srcip6(cmd, *av)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
}
break;
case TOK_DSTIP6:
NEED1("missing destination IP6");
if (add_dstip6(cmd, *av)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
}
break;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
case TOK_SRCPORT:
NEED1("missing source port");
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(*av, "any") == 0 ||
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
add_ports(cmd, *av, proto, O_IP_SRCPORT)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
} else
errx(EX_DATAERR, "invalid source port %s", *av);
break;
case TOK_DSTPORT:
NEED1("missing destination port");
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(*av, "any") == 0 ||
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
add_ports(cmd, *av, proto, O_IP_DSTPORT)) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
} else
errx(EX_DATAERR, "invalid destination port %s",
*av);
break;
case TOK_MAC:
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (add_mac(cmd, av))
av += 2;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
break;
case TOK_MACTYPE:
NEED1("missing mac type");
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (!add_mactype(cmd, *av))
o Pass a correct argument to printf(3). PR: bin/51750 Submitted by: Vasil Dimov <vd@datamax.bg> MFC after: 2 weeks
2003-06-16 09:44:53 +00:00
errx(EX_DATAERR, "invalid mac type %s", *av);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Major cleanup of the parser and printing routines in an attempt to render the syntax less ambiguous. Now rules can be in one of these two forms <action> <protocol> from <src> to <dst> [options] <action> MAC dst-mac src-mac mac-type [options] however you can now specify MAC and IP header fields as options e.g. ipfw add allow all from any to any mac-type arp ipfw add allow all from any to any { dst-ip me or src-ip me } which makes complex expressions a lot easier to write and parse. The "all from any to any" part is there just for backward compatibility. Manpage updated accordingly.
2002-08-19 04:52:15 +00:00
break;
Add a 'verrevpath' option that verifies the interface that a packet comes in on is the same interface that we would route out of to get to the packet's source address. Essentially automates an anti-spoofing check using the information in the routing table. Experimental. The usage and rule format for the feature may still be subject to change.
2003-03-15 01:13:00 +00:00
case TOK_VERREVPATH:
fill_cmd(cmd, O_VERREVPATH, 0, 0);
break;
remove extra whitespace and blank lines
2003-06-27 17:18:14 +00:00
Add the option versrcreach to verify that a valid route to the source address of a packet exists in the routing table. The default route is ignored because it would match everything and render the check pointless. This option is very useful for routers with a complete view of the Internet (BGP) in the routing table to reject packets with spoofed or unrouteable source addresses. Example: ipfw add 1000 deny ip from any to any not versrcreach also known in Cisco-speak as: ip verify unicast source reachable-via any Reviewed by: luigi
2004-04-23 14:28:38 +00:00
case TOK_VERSRCREACH:
fill_cmd(cmd, O_VERSRCREACH, 0, 0);
break;
New ipfw option "antispoof": For incoming packets, the packet's source address is checked if it belongs to a directly connected network. If the network is directly connected, then the interface the packet came on in is compared to the interface the network is connected to. When incoming interface and directly connected interface are not the same, the packet does not match. Usage example: ipfw add deny ip from any to any not antispoof in Manpage education by: ru
2004-08-09 16:12:10 +00:00
case TOK_ANTISPOOF:
fill_cmd(cmd, O_ANTISPOOF, 0, 0);
break;
Implement the 'ipsec' option to match packets coming out of an ipsec tunnel. Should work with both regular and fast ipsec (mutually exclusive). See manpage for more details. Submitted by: Ari Suutari (ari.suutari@syncrontech.com) Revised by: sam MFC after: 1 week
2003-07-04 21:42:32 +00:00
case TOK_IPSEC:
fill_cmd(cmd, O_IPSEC, 0, 0);
break;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
case TOK_IPV6:
fill_cmd(cmd, O_IP6, 0, 0);
break;
Add support for IPv4 only rules to IPFW2 now that it supports IPv6 as well. This is the last requirement before we can retire ip6fw. Reviewed by: dwhite, brooks(earlier version) Submitted by: dwhite (manpage) Silence from: -ipfw
2005-06-03 01:10:28 +00:00
case TOK_IPV4:
fill_cmd(cmd, O_IP4, 0, 0);
break;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
case TOK_EXT6HDR:
fill_ext6hdr( cmd, *av );
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
break;
case TOK_FLOWID:
if (proto != IPPROTO_IPV6 )
errx( EX_USAGE, "flow-id filter is active "
"only for ipv6 protocol\n");
fill_flow6( (ipfw_insn_u32 *) cmd, *av );
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add IPv6 support to IPFW and Dummynet. Submitted by: Mariano Tortoriello and Raffaele De Lorenzo (via luigi)
2005-04-18 18:35:05 +00:00
break;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
case TOK_COMMENT:
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
fill_comment(cmd, av);
av[0]=NULL;
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
break;
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
case TOK_TAGGED:
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (av[0] && strpbrk(*av, "-,")) {
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
if (!add_ports(cmd, *av, 0, O_TAGGED))
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
errx(EX_DATAERR, "tagged: invalid tag"
" list: %s", *av);
}
else {
uint16_t tag;
Several ipfw options and actions use a 16-bit argument to indicate pipes, queues, tags, rule numbers and so on. These are all different namespaces, and the only thing they have in common is the fact they use a 16-bit slot to represent the argument. There is some confusion in the code, mostly for historical reasons, on how the values 0 and 65535 should be used. At the moment, 0 is forbidden almost everywhere, while 65535 is used to represent a 'tablearg' argument, i.e. the result of the most recent table() lookup. For now, try to use explicit constants for the min and max allowed values, and do not overload the default rule number for that. Also, make the MTAG_IPFW declaration only visible to the kernel. NOTE: I think the issue needs to be revisited before 8.0 is out: the 2^16 namespace limit for rule numbers and pipe/queue is annoying, and we can easily bump the limit to 2^32 which gives a lot more flexibility in partitioning the namespace. MFC after: 5 days
2009-06-05 16:16:07 +00:00
GET_UINT_ARG(tag, IPFW_ARG_MIN, IPFW_ARG_MAX,
Use IPFW_DEFAULT_RULE instead of hardcoded value since now it is available. MFC after: 5 days.
2008-09-06 17:23:37 +00:00
TOK_TAGGED, rule_options);
Add support of 'tablearg' feature for: - 'tag' & 'untag' action parameters. - 'tagged' & 'limit' rule options. Rule examples: pipe 1 tag tablearg ip from table(1) to any allow ip from any to table(2) tagged tablearg allow tcp from table(3) to any 25 setup limit src-addr tablearg sbin/ipfw/ipfw2.c: 1) new macros GET_UINT_ARG - support of 'tablearg' keyword, argument range checking. PRINT_UINT_ARG - support of 'tablearg' keyword. 2) strtoport(): do not silently truncate/accept invalid port list expressions like: '1,2-abc' or '1,2-3-4' or '1,2-3x4'. style(9) cleanup. Approved by: glebius (mentor) MFC after: 1 month
2006-06-15 09:39:22 +00:00
fill_cmd(cmd, O_TAGGED, 0, tag);
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
}
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
break;
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
case TOK_FIB:
NEED1("fib requires fib number");
fill_cmd(cmd, O_FIB, 0, strtoul(*av, NULL, 0));
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
break;
implement a new match option, lookup {dst-ip|src-ip|dst-port|src-port|uid|jail} N which searches the specified field in table N and sets tablearg accordingly. With dst-ip or src-ip the option replicates two existing options. When used with other arguments, the option can be useful to quickly dispatch traffic based on other fields. Work supported by the Onelab project. MFC after: 1 week
2009-12-15 09:46:27 +00:00
case TOK_LOOKUP: {
ipfw_insn_u32 *c = (ipfw_insn_u32 *)cmd;
char *p;
int j;
Implement "lookup dscp N" which does a lookup of the DSCP (top 6 bits of ip->ip_tos) in a table. This can be useful to direct traffic to different pipes/queues according to the DSCP of the packet, as follows: ipfw add 100 queue tablearg lookup dscp 3 // table 3 maps dscp->queue This change is a no-op (but harmless) until the two-line kernel side is committed, which will happen shortly.
2010-03-15 15:43:35 +00:00
if (!av[0] || !av[1])
implement a new match option, lookup {dst-ip|src-ip|dst-port|src-port|uid|jail} N which searches the specified field in table N and sets tablearg accordingly. With dst-ip or src-ip the option replicates two existing options. When used with other arguments, the option can be useful to quickly dispatch traffic based on other fields. Work supported by the Onelab project. MFC after: 1 week
2009-12-15 09:46:27 +00:00
errx(EX_USAGE, "format: lookup argument tablenum");
cmd->opcode = O_IP_DST_LOOKUP;
cmd->len |= F_INSN_SIZE(ipfw_insn) + 2;
i = match_token(rule_options, *av);
for (j = 0; lookup_key[j] >= 0 ; j++) {
if (i == lookup_key[j])
break;
}
if (lookup_key[j] <= 0)
errx(EX_USAGE, "format: cannot lookup on %s", *av);
c->d[1] = j; // i converted to option
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
implement a new match option, lookup {dst-ip|src-ip|dst-port|src-port|uid|jail} N which searches the specified field in table N and sets tablearg accordingly. With dst-ip or src-ip the option replicates two existing options. When used with other arguments, the option can be useful to quickly dispatch traffic based on other fields. Work supported by the Onelab project. MFC after: 1 week
2009-12-15 09:46:27 +00:00
cmd->arg1 = strtoul(*av, &p, 0);
if (p && *p)
errx(EX_USAGE, "format: lookup argument tablenum");
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
av++;
implement a new match option, lookup {dst-ip|src-ip|dst-port|src-port|uid|jail} N which searches the specified field in table N and sets tablearg accordingly. With dst-ip or src-ip the option replicates two existing options. When used with other arguments, the option can be useful to quickly dispatch traffic based on other fields. Work supported by the Onelab project. MFC after: 1 week
2009-12-15 09:46:27 +00:00
}
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
default:
errx(EX_USAGE, "unrecognised option [%d] %s\n", i, s);
}
if (F_LEN(cmd) > 0) { /* prepare to advance */
prev = cmd;
cmd = next_cmd(cmd);
}
}
done:
/*
* Now copy stuff into the rule.
* If we have a keep-state option, the first instruction
* must be a PROBE_STATE (which is generated here).
* If we have a LOG option, it was stored as the first command,
* and now must be moved to the top of the action part.
*/
dst = (ipfw_insn *)rule->cmd;
Fix a kernel panic with rules of the type prob 0.5 pipe NN .... due to the generation of an invalid ipfw instruction sequence. No ABI change, but you need to upgrade /sbin/ipfw to generate the correct code. Approved by: re
2002-11-26 19:58:12 +00:00
/*
* First thing to write into the command stream is the match probability.
*/
if (match_prob != 1) { /* 1 means always match */
dst->opcode = O_PROB;
dst->len = 2;
*((int32_t *)(dst+1)) = (int32_t)(match_prob * 0x7fffffff);
dst += dst->len;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* generate O_PROBE_STATE if necessary
*/
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
if (have_state && have_state->opcode != O_CHECK_STATE) {
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
fill_cmd(dst, O_PROBE_STATE, 0, 0);
dst = next_cmd(dst);
}
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
/* copy all commands but O_LOG, O_KEEP_STATE, O_LIMIT, O_ALTQ, O_TAG */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
for (src = (ipfw_insn *)cmdbuf; src != cmd; src += i) {
i = F_LEN(src);
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
switch (src->opcode) {
case O_LOG:
case O_KEEP_STATE:
case O_LIMIT:
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
case O_ALTQ:
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
case O_TAG:
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
break;
default:
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
bcopy(src, dst, i * sizeof(uint32_t));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
dst += i;
}
}
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
/*
* put back the have_state command as last opcode
*/
Fix generation of check-state rules, which i broke in last commit.
2002-08-04 05:16:19 +00:00
if (have_state && have_state->opcode != O_CHECK_STATE) {
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
i = F_LEN(have_state);
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
bcopy(have_state, dst, i * sizeof(uint32_t));
Two bugfixes: + the header file contains two different opcodes (O_IPOPTS and O_IPOPT) for what is the same thing, and sure enough i used one in the kernel and the other one in userland. Be consistent! + "keep-state" and "limit" must be the last match pattern in a rule, so no matter how you enter them move them to the end of the rule.
2002-07-31 22:31:47 +00:00
dst += i;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* start action section
*/
rule->act_ofs = dst - rule->cmd;
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
/* put back O_LOG, O_ALTQ, O_TAG if necessary */
Add to IPFW the ability to do ALTQ classification/tagging.
2004-10-03 00:17:46 +00:00
if (have_log) {
i = F_LEN(have_log);
bcopy(have_log, dst, i * sizeof(uint32_t));
dst += i;
}
if (have_altq) {
i = F_LEN(have_altq);
bcopy(have_altq, dst, i * sizeof(uint32_t));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
dst += i;
}
Implement internal (i.e. inside kernel) packet tagging using mbuf_tags(9). Since tags are kept while packet resides in kernelspace, it's possible to use other kernel facilities (like netgraph nodes) for altering those tags. Submitted by: Andrey Elsukov <bu7cher at yandex dot ru> Submitted by: Vadim Goncharov <vadimnuclight at tpu dot ru> Approved by: glebius (mentor) Idea from: OpenBSD PF MFC after: 1 month
2006-05-24 13:09:55 +00:00
if (have_tag) {
i = F_LEN(have_tag);
bcopy(have_tag, dst, i * sizeof(uint32_t));
dst += i;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* copy all other actions
*/
for (src = (ipfw_insn *)actbuf; src != action; src += i) {
i = F_LEN(src);
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
bcopy(src, dst, i * sizeof(uint32_t));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
dst += i;
}
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
rule->cmd_len = (uint32_t *)dst - (uint32_t *)(rule->cmd);
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
i = (char *)dst - (char *)rule;
Apply a bandaid to get this working on sparc64 again; the introduction of do_cmd() broke things, because this function assumes that a socklen_t is large enough to hold a pointer. A real solution to this problem would be a rewrite of do_cmd() to treat the optlen parameter consistently and not use it to carry a pointer or integer dependent on the context.
2003-09-04 15:57:37 +00:00
if (do_cmd(IP_FW_ADD, rule, (uintptr_t)&i) == -1)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
err(EX_UNAVAILABLE, "getsockopt(%s)", "IP_FW_ADD");
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (!co.do_quiet)
In random order: * make the code compile with WARNS=5 (at least on i386), mostly by adding 'const' specifier and replacing "void *" with "char *" in places where pointer arithmetic was used. This also spotted a few places where invalid tests (e.g. uint < 0) were used. * support ranges in "list" and "show" commands. Now you can say ipfw show 100-1000 4000-8000 which is very convenient when you have large rulesets. * implement comments in ipfw commands. These are implemented in the kernel as O_NOP commands (which always match) whose body contains the comment string. In userland, a comment is a C++-style comment: ipfw add allow ip from me to any // i can talk to everybody The choice of '//' versus '#' is somewhat arbitrary, but because the preprocessor/readfile part of ipfw used to strip away '#', I did not want to change this behaviour. If a rule only contains a comment ipfw add 1000 // this rule is just a comment then it is stored as a 'count' rule (this is also to remind the user that scanning through a rule is expensive). * improve handling of flags (still to be completed). ipfw_main() was written thinking of 'one rule per ipfw invocation', and so flags are set and never cleared. With readfile/preprocessor support, this changes and certain flags should be reset on each line. For the time being, only fix handling of '-a' which differentiates the "list" and "show" commands. * rework the preprocessor support -- ipfw_main() already had most of the parsing code, so i have moved in there the only missing bit (stripping away '#' and comments) and removed the parsing from ipfw_readfile(). Also, add some more options (such as -c, -N, -S) to the readfile section. MFC after: 3 days
2003-07-12 06:53:16 +00:00
show_ipfw(rule, 0, 0);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
/*
* clear the counters or the log counters.
*/
void
ipfw_zero(int ac, char *av[], int optname /* 0 = IP_FW_ZERO, 1 = IP_FW_RESETLOG */)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
uint32_t arg, saved_arg;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int failed = EX_OK;
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
char const *errstr;
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
char const *name = optname ? "RESETLOG" : "ZERO";
optname = optname ? IP_FW_RESETLOG : IP_FW_ZERO;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
av++; ac--;
if (!ac) {
/* clear all entries */
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
if (do_cmd(optname, NULL, 0) < 0)
err(EX_UNAVAILABLE, "setsockopt(IP_FW_%s)", name);
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (!co.do_quiet)
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
printf("%s.\n", optname == IP_FW_ZERO ?
"Accounting cleared":"Logging counts reset");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return;
}
while (ac) {
/* Rule number */
if (isdigit(**av)) {
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
arg = strtonum(*av, 0, 0xffff, &errstr);
if (errstr)
errx(EX_DATAERR,
"invalid rule number %s\n", *av);
saved_arg = arg;
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set)
arg |= (1 << 24) | ((co.use_set - 1) << 16);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
av++;
ac--;
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
if (do_cmd(optname, &arg, sizeof(arg))) {
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
warn("rule %u: setsockopt(IP_FW_%s)",
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
saved_arg, name);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
failed = EX_UNAVAILABLE;
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
} else if (!co.do_quiet)
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
printf("Entry %d %s.\n", saved_arg,
A bunch of changes (mostly syntactic sugar, all backward compatible): * Make the addr-set size optional (defaults to /24) You can now write 1.2.3.0/24{56-80} or 1.2.3.0{56-80} Also make the parser more strict. * Support a new format for the list of addresses: 1.2.3.4,5.6.7.8/30,9.10.11.12/22,12.12.12.13, ... which exploits the new capabilities of O_IP_SRC_MASK/O_IP_DST_MASK * Allow spaces after commas to make lists of addresses more readable. 1.2.3.4, 5.6.7.8/30, 9.10.11.12/22, 12.12.12.13, ... * ipfw will now accept full commands as a single argument and strip extra leading/trailing whitespace as below: ipfw "-q add allow ip from 1.2.3.4 to 5.6.7.8, 9.10.11.23 " This should help in moving the body of ipfw into a library that user programs can invoke. * Cleanup some comments and data structures. * Do not print rule counters for dynamic rules with ipfw -d list (PR 51182) * Improve 'ipfw -h' output (PR 46785) * Add a '-n' flag to test the syntax of commands without actually calling [gs]etsockopt() (PR 44238) * Support the '-n' flag also with the preprocessors; Manpage commit to follow. MFC after: 3 days
2003-07-08 07:52:47 +00:00
optname == IP_FW_ZERO ?
"cleared" : "logging count reset");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
} else {
errx(EX_USAGE, "invalid rule number ``%s''", *av);
}
}
if (failed != EX_OK)
exit(failed);
}
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
void
ipfw_flush(int force)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
int cmd = co.do_pipe ? IP_DUMMYNET_FLUSH : IP_FW_FLUSH;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (!force && !co.do_quiet) { /* need to ask user */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
int c;
printf("Are you sure? [yn] ");
fflush(stdout);
do {
c = toupper(getc(stdin));
while (c != '\n' && getc(stdin) != '\n')
if (feof(stdin))
return; /* and do not flush */
} while (c != 'Y' && c != 'N');
printf("\n");
if (c == 'N') /* user said no */
return;
}
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
if (co.do_pipe) {
dummynet_flush();
return;
}
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
/* `ipfw set N flush` - is the same that `ipfw delete set N` */
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.use_set) {
uint32_t arg = ((co.use_set - 1) & 0xffff) | (1 << 24);
o Make ipfw set more robust -- now it is possible: - to show a specific set: ipfw set 3 show - to delete rules from the set: ipfw set 9 delete 100 200 300 - to flush the set: ipfw set 4 flush - to reset rules counters in the set: ipfw set 1 zero PR: kern/113388 Submitted by: Andrey V. Elsukov Approved by: re (kensmith) MFC after: 6 weeks
2007-06-18 17:52:37 +00:00
if (do_cmd(IP_FW_DEL, &arg, sizeof(arg)) < 0)
err(EX_UNAVAILABLE, "setsockopt(IP_FW_DEL)");
} else if (do_cmd(cmd, NULL, 0) < 0)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
err(EX_UNAVAILABLE, "setsockopt(IP_%s_FLUSH)",
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
co.do_pipe ? "DUMMYNET" : "FW");
if (!co.do_quiet)
printf("Flushed all %s.\n", co.do_pipe ? "pipes" : "rules");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
ccept of empty lines when reading from a file (this fixes a bug introduced in the latest commits). Also: * update the 'ipfw -h' output; * allow rules of the form "100 add allow ..." i.e. with the index first. (requested by Paul Richards). This was an undocumented ipfw1 behaviour, and it is left undocumented. and minor code cleanups.
2003-07-14 08:39:49 +00:00
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
static void table_list(ipfw_table_entry ent, int need_header);
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
/*
* This one handles all table-related commands
* ipfw table N add addr[/masklen] [value]
* ipfw table N delete addr[/masklen]
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
* ipfw table {N | all} flush
* ipfw table {N | all} list
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
*/
Start splitting the monster file in smaller blocks. In this episode: - introduce a common header with a minimal set of common definitions; - bring the main() function and options parser in main.c - rename the main functions with an ipfw_ prefix No code changes except for the introduction of a global variable, resvd_set_number, which stores the RESVD_SET value from ip_fw.h and is used to remove the dependency of main.c from ip_fw.h (and the subtree of dependencies) for just a single constant.
2009-01-27 10:18:55 +00:00
void
ipfw_table_handler(int ac, char *av[])
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
{
ipfw_table_entry ent;
int do_add;
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
int is_all;
Add the check of the table number.
2008-09-21 21:46:56 +00:00
size_t len;
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
char *p;
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
uint32_t a;
Add the check of the table number.
2008-09-21 21:46:56 +00:00
uint32_t tables_max;
Unbreak the build.
2008-09-22 04:12:27 +00:00
len = sizeof(tables_max);
Add the check of the table number.
2008-09-21 21:46:56 +00:00
if (sysctlbyname("net.inet.ip.fw.tables_max", &tables_max, &len,
NULL, 0) == -1) {
#ifdef IPFW_TABLES_MAX
warn("Warn: Failed to get the max tables number via sysctl. "
"Using the compiled in defaults. \nThe reason was");
tables_max = IPFW_TABLES_MAX;
#else
errx(1, "Failed sysctlbyname(\"net.inet.ip.fw.tables_max\")");
#endif
}
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
ac--; av++;
if (ac && isdigit(**av)) {
ent.tbl = atoi(*av);
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
is_all = 0;
ac--; av++;
} else if (ac && _substrcmp(*av, "all") == 0) {
ent.tbl = 0;
is_all = 1;
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
ac--; av++;
} else
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
errx(EX_USAGE, "table number or 'all' keyword required");
Add the check of the table number.
2008-09-21 21:46:56 +00:00
if (ent.tbl >= tables_max)
errx(EX_USAGE, "The table number exceeds the maximum allowed "
"value (%d)", tables_max - 1);
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
NEED1("table needs command");
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
if (is_all && _substrcmp(*av, "list") != 0
&& _substrcmp(*av, "flush") != 0)
errx(EX_USAGE, "table number required");
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
if (_substrcmp(*av, "add") == 0 ||
_substrcmp(*av, "delete") == 0) {
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
do_add = **av == 'a';
ac--; av++;
if (!ac)
errx(EX_USAGE, "IP address required");
p = strchr(*av, '/');
if (p) {
*p++ = '\0';
ent.masklen = atoi(p);
if (ent.masklen > 32)
errx(EX_DATAERR, "bad width ``%s''", p);
} else
ent.masklen = 32;
if (lookup_host(*av, (struct in_addr *)&ent.addr) != 0)
Fixed a bug spotted by compiling with -Wall.
2004-06-10 07:51:11 +00:00
errx(EX_NOHOST, "hostname ``%s'' unknown", *av);
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
ac--; av++;
Allow ipfw to forward to a destination that is specified by a table. for example: fwd tablearg ip from any to table(1) where table 1 has entries of the form: 1.1.1.0/24 10.2.3.4 208.23.2.0/24 router2 This allows trivial implementation of a secondary routing table implemented in the firewall layer. I expect more work (under discussion with Glebius) to follow this to clean up some of the messy parts of ipfw related to tables. Reviewed by: Glebius MFC after: 1 month
2006-08-17 22:49:50 +00:00
if (do_add && ac) {
unsigned int tval;
/* isdigit is a bit of a hack here.. */
if (strchr(*av, (int)'.') == NULL && isdigit(**av)) {
ent.value = strtoul(*av, NULL, 0);
} else {
if (lookup_host(*av, (struct in_addr *)&tval) == 0) {
/* The value must be stored in host order *
* so that the values < 65k can be distinguished */
ent.value = ntohl(tval);
} else {
errx(EX_NOHOST, "hostname ``%s'' unknown", *av);
}
}
} else
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
ent.value = 0;
if (do_cmd(do_add ? IP_FW_TABLE_ADD : IP_FW_TABLE_DEL,
Amazing.. two screwups in one commit. I'm piling on thise pointy hats on top of each other. At least they nest..
2006-03-31 12:54:17 +00:00
&ent, sizeof(ent)) < 0) {
Stop ipfw from aborting when asked to delete a table entry that doesn't exist or add one that is already present, if the -q flag is set. Useful for "ipfw -q /dev/stdin" when the command above is invoked from something like python or TCL to feed commands down the throat of ipfw. MFC in: 1 week
2006-02-14 03:10:29 +00:00
/* If running silent, don't bomb out on these errors. */
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (!(co.do_quiet && (errno == (do_add ? EEXIST : ESRCH))))
Stop ipfw from aborting when asked to delete a table entry that doesn't exist or add one that is already present, if the -q flag is set. Useful for "ipfw -q /dev/stdin" when the command above is invoked from something like python or TCL to feed commands down the throat of ipfw. MFC in: 1 week
2006-02-14 03:10:29 +00:00
err(EX_OSERR, "setsockopt(IP_FW_TABLE_%s)",
do_add ? "ADD" : "DEL");
/* In silent mode, react to a failed add by deleting */
I can't believe that no-one noticed that I broke ipfw table del for over a month! put {} around if clause with multiple statements
2006-03-31 12:40:09 +00:00
if (do_add) {
Stop ipfw from aborting when asked to delete a table entry that doesn't exist or add one that is already present, if the -q flag is set. Useful for "ipfw -q /dev/stdin" when the command above is invoked from something like python or TCL to feed commands down the throat of ipfw. MFC in: 1 week
2006-02-14 03:10:29 +00:00
do_cmd(IP_FW_TABLE_DEL, &ent, sizeof(ent));
if (do_cmd(IP_FW_TABLE_ADD,
&ent, sizeof(ent)) < 0)
err(EX_OSERR,
"setsockopt(IP_FW_TABLE_ADD)");
I can't believe that no-one noticed that I broke ipfw table del for over a month! put {} around if clause with multiple statements
2006-03-31 12:40:09 +00:00
}
Amazing.. two screwups in one commit. I'm piling on thise pointy hats on top of each other. At least they nest..
2006-03-31 12:54:17 +00:00
}
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "flush") == 0) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
a = is_all ? tables_max : (uint32_t)(ent.tbl + 1);
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
do {
if (do_cmd(IP_FW_TABLE_FLUSH, &ent.tbl,
sizeof(ent.tbl)) < 0)
err(EX_OSERR, "setsockopt(IP_FW_TABLE_FLUSH)");
} while (++ent.tbl < a);
Deprecate unmaintainable uses of strncmp to implement abbreviations. This commit replaces those with two new functions that simplify the code and produce warnings that the syntax is deprecated. A small number of sensible abbreviations may be explicitly added based on user feedback. There were previously three types of strncmp use in ipfw: - Most commonly, strncmp(av, "string", sizeof(av)) was used to allow av to match string or any shortened form of it. I have replaced this with a new function _substrcmp(av, "string") which returns 0 if av is a substring of "string", but emits a warning if av is not exactly "string". - The next type was two instances of strncmp(av, "by", 2) which allowed the abbreviation of bytes to "by", "byt", etc. Unfortunately, it also supported "bykHUygh&*g&*7*ui". I added a second new function _substrcmp2(av, "by", "bytes") which acts like the strncmp did, but complains if the user doesn't spell out the word "bytes". - There is also one correct use of strncmp to match "table(" which might have another token after it without a space. Since I changed all the lines anyway, I also fixed the treatment of strncmp's return as a boolean in many cases. I also modified a few strcmp cases as well to be fully consistent.
2005-01-15 01:46:41 +00:00
} else if (_substrcmp(*av, "list") == 0) {
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
a = is_all ? tables_max : (uint32_t)(ent.tbl + 1);
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
do {
Fix the build. Noted by: ganbold@
2008-09-27 15:58:54 +00:00
table_list(ent, is_all);
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
} while (++ent.tbl < a);
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
} else
errx(EX_USAGE, "invalid table command %s", *av);
}
Do not do the useless job for an empty table. MFC after: 1 month
2008-09-20 15:54:22 +00:00
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
static void
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
table_list(ipfw_table_entry ent, int need_header)
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
{
ipfw_table *tbl;
socklen_t l;
uint32_t a;
Do not do the useless job for an empty table. MFC after: 1 month
2008-09-20 15:54:22 +00:00
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
a = ent.tbl;
l = sizeof(a);
if (do_cmd(IP_FW_TABLE_GETSIZE, &a, (uintptr_t)&l) < 0)
err(EX_OSERR, "getsockopt(IP_FW_TABLE_GETSIZE)");
/* If a is zero we have nothing to do, the table is empty. */
if (a == 0)
return;
l = sizeof(*tbl) + a * sizeof(ipfw_table_entry);
wrap all malloc/calloc/realloc calls so they exit on failure without having to check in each place. Remove an wrong strdup from previous commit.
2009-01-26 14:26:35 +00:00
tbl = safe_calloc(1, l);
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
tbl->tbl = ent.tbl;
if (do_cmd(IP_FW_TABLE_LIST, tbl, (uintptr_t)&l) < 0)
err(EX_OSERR, "getsockopt(IP_FW_TABLE_LIST)");
Add keyword all in addtion to the table number for the 'list' and the 'flush' actions on tables. Part of PR: 127058. PR: 127058 (based on) MFC after: 1 month
2008-09-27 14:30:34 +00:00
if (tbl->cnt && need_header)
printf("---table(%d)---\n", tbl->tbl);
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
for (a = 0; a < tbl->cnt; a++) {
unsigned int tval;
tval = tbl->ent[a].value;
put all options in a single struct, and document them. This will allow us to easily restore the original values when processing commands from a file (where each individual line can have its own options).
2009-01-27 09:06:25 +00:00
if (co.do_value_as_ip) {
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
char tbuf[128];
strncpy(tbuf, inet_ntoa(*(struct in_addr *)
comply with style police Submitted by: ru MFC after: 1 month
2006-08-18 22:36:05 +00:00
&tbl->ent[a].addr), 127);
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
/* inet_ntoa expects network order */
tval = htonl(tval);
printf("%s/%u %s\n", tbuf, tbl->ent[a].masklen,
inet_ntoa(*(struct in_addr *)&tval));
} else {
printf("%s/%u %u\n",
inet_ntoa(*(struct in_addr *)&tbl->ent[a].addr),
tbl->ent[a].masklen, tval);
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
}
Move table list to a separate function.
2008-09-21 12:54:09 +00:00
}
free(tbl);
Introduce a new feature to IPFW2: lookup tables. These are useful for handling large sparse address sets. Initial implementation by Vsevolod Lobko <seva@ip.net.ua>, refined by me. MFC after: 1 week
2004-06-09 20:10:38 +00:00
}
Reference in New Issue Copy Permalink