d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
306fe283a1
freebsd-dev/sbin/ipfw/ipfw2.c

3391 lines
76 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
/*
* Copyright (c) 2002 Luigi Rizzo
* 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
*
* $FreeBSD$
*/
#include <sys/param.h>
#include <sys/mbuf.h>
#include <sys/socket.h>
#include <sys/sockio.h>
#include <sys/sysctl.h>
#include <sys/time.h>
#include <sys/wait.h>
#include <ctype.h>
#include <err.h>
#include <errno.h>
#include <grp.h>
#include <limits.h>
#include <netdb.h>
#include <pwd.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <string.h>
Fixed some world breakage caused by not updating clients when <timeconv.h> was split off from <time.h>. This became fatal here when -Werror was reenabled.
2002-07-08 19:49:52 +00:00
#include <timeconv.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 <unistd.h>
#include <sysexits.h>
#include <net/if.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/ip_icmp.h>
#include <netinet/ip_fw.h>
#include <net/route.h> /* def. of struct route */
#include <netinet/ip_dummynet.h>
#include <netinet/tcp.h>
#include <arpa/inet.h>
int s, /* main RAW socket */
do_resolv, /* Would try to resolve all */
do_acct, /* Show packet/byte count */
do_time, /* Show time stamps */
do_quiet, /* Be quiet in add and flush */
do_force, /* Don't ask for confirmation */
do_pipe, /* this cmd refers to a pipe */
do_sort, /* field to sort results (0 = no) */
do_dynamic, /* display dynamic rules */
do_expired, /* display expired dynamic rules */
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
show_sets, /* display rule sets */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
verbose;
#define IP_MASK_ALL 0xffffffff
/*
* structure to hold flag names and associated values to be
* set in the appropriate masks.
* A NULL string terminates the array.
* Often, an element with 0 value contains an error string.
*
*/
struct _s_x {
char *s;
int x;
};
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},
{ "congestion", IPTOS_CE},
{ "ecntransport", IPTOS_ECT},
{ "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 }
};
static void show_usage(void);
enum tokens {
TOK_NULL=0,
TOK_OR,
TOK_NOT,
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,
TOK_ENDBRACE,
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
TOK_ACCEPT,
TOK_COUNT,
TOK_PIPE,
TOK_QUEUE,
TOK_DIVERT,
TOK_TEE,
TOK_FORWARD,
TOK_SKIPTO,
TOK_DENY,
TOK_REJECT,
TOK_RESET,
TOK_UNREACH,
TOK_CHECKSTATE,
TOK_UID,
TOK_GID,
TOK_IN,
TOK_LIMIT,
TOK_KEEPSTATE,
TOK_LAYER2,
TOK_OUT,
TOK_XMIT,
TOK_RECV,
TOK_VIA,
TOK_FRAG,
TOK_IPOPTS,
TOK_IPLEN,
TOK_IPID,
TOK_IPPRECEDENCE,
TOK_IPTOS,
TOK_IPTTL,
TOK_IPVER,
TOK_ESTAB,
TOK_SETUP,
TOK_TCPFLAGS,
TOK_TCPOPTS,
TOK_TCPSEQ,
TOK_TCPACK,
TOK_TCPWIN,
TOK_ICMPTYPES,
TOK_PLR,
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
TOK_NOERROR,
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
TOK_BUCKETS,
TOK_DSTIP,
TOK_SRCIP,
TOK_DSTPORT,
TOK_SRCPORT,
TOK_ALL,
TOK_MASK,
TOK_BW,
TOK_DELAY,
TOK_RED,
TOK_GRED,
TOK_DROPTAIL,
TOK_PROTO,
TOK_WEIGHT,
};
struct _s_x dummynet_params[] = {
{ "plr", TOK_PLR },
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
{ "noerror", TOK_NOERROR },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "buckets", TOK_BUCKETS },
{ "dst-ip", TOK_DSTIP },
{ "src-ip", TOK_SRCIP },
{ "dst-port", TOK_DSTPORT },
{ "src-port", TOK_SRCPORT },
{ "proto", TOK_PROTO },
{ "weight", TOK_WEIGHT },
{ "all", TOK_ALL },
{ "mask", TOK_MASK },
{ "droptail", TOK_DROPTAIL },
{ "red", TOK_RED },
{ "gred", TOK_GRED },
{ "bw", TOK_BW },
{ "bandwidth", TOK_BW },
{ "delay", TOK_DELAY },
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
{ "pipe", TOK_PIPE },
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ "queue", TOK_QUEUE },
{ "dummynet-params", TOK_NULL },
{ NULL, 0 }
};
struct _s_x rule_actions[] = {
{ "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 },
{ "fwd", TOK_FORWARD },
{ "forward", TOK_FORWARD },
{ "skipto", TOK_SKIPTO },
{ "deny", TOK_DENY },
{ "drop", TOK_DENY },
{ "reject", TOK_REJECT },
{ "reset", TOK_RESET },
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 },
{ NULL, TOK_NULL },
{ NULL, 0 }
};
struct _s_x rule_options[] = {
{ "uid", TOK_UID },
{ "gid", TOK_GID },
{ "in", TOK_IN },
{ "limit", TOK_LIMIT },
{ "keep-state", TOK_KEEPSTATE },
{ "bridged", TOK_LAYER2 },
{ "layer2", TOK_LAYER2 },
{ "out", TOK_OUT },
{ "xmit", TOK_XMIT },
{ "recv", TOK_RECV },
{ "via", TOK_VIA },
{ "fragment", TOK_FRAG },
{ "frag", TOK_FRAG },
{ "ipoptions", TOK_IPOPTS },
{ "ipopts", TOK_IPOPTS },
{ "iplen", TOK_IPLEN },
{ "ipid", TOK_IPID },
{ "ipprecedence", TOK_IPPRECEDENCE },
{ "iptos", TOK_IPTOS },
{ "ipttl", TOK_IPTTL },
{ "ipversion", TOK_IPVER },
{ "ipver", TOK_IPVER },
{ "estab", TOK_ESTAB },
{ "established", TOK_ESTAB },
{ "setup", TOK_SETUP },
{ "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 },
{ "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 */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{ NULL, TOK_NULL },
{ NULL, 0 }
};
/**
* match_token takes a table and a string, returns the value associated
* with the string (0 meaning an error in most cases)
*/
static int
match_token(struct _s_x *table, char *string)
{
struct _s_x *pt;
int i = strlen(string);
for (pt = table ; i && pt->s != NULL ; pt++)
if (strlen(pt->s) == i && !bcmp(string, pt->s, i))
return pt->x;
return -1;
};
static char *
match_value(struct _s_x *p, u_int32_t value)
{
for (; p->s != NULL; p++)
if (p->x == value)
return p->s;
return NULL;
}
/*
* prints one port, symbolic or numeric
*/
static void
print_port(int proto, u_int16_t port)
{
if (proto == IPPROTO_ETHERTYPE) {
char *s;
if (do_resolv && (s = match_value(ether_types, port)) )
printf("%s", s);
else
printf("0x%04x", port);
} else {
struct servent *se = NULL;
if (do_resolv) {
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);
}
}
/*
* print the values in a list of ports
* XXX todo: add support for mask.
*/
static void
print_newports(ipfw_insn_u16 *cmd, int proto)
{
u_int16_t *p = cmd->ports;
int i;
char *sep= " ";
if (cmd->o.len & F_NOT)
printf(" not");
for (i = F_LEN((ipfw_insn *)cmd) - 1; i > 0; i--, p += 2) {
printf(sep);
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 */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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 == '\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 */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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++;
buf = malloc(s1 - s + 1);
if (buf == NULL)
return 0;
/*
* copy into a buffer skipping backslashes
*/
for (p = s, i = 0; p != s1 ; p++)
if ( *p != '\\')
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
}
/*
* fill the body of the command with the list of port ranges.
* At the moment it only understands numeric ranges.
*/
static int
fill_newports(ipfw_insn_u16 *cmd, char *av, int proto)
{
u_int16_t *p = cmd->ports;
int i = 0;
for (; *av ; i++, p +=2 ) {
u_int16_t a, b;
char *s;
a = strtoport(av, &s, 0, proto);
if (s == av) /* no parameter */
break;
if (*s == '-') { /* a range */
av = s+1;
b = strtoport(av, &s, 0, proto);
if (s == av) /* no parameter */
break;
p[0] = a;
p[1] = b;
} else if (*s == ',' || *s == '\0' ) {
p[0] = p[1] = a;
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 { /* invalid separator */
errx(EX_DATAERR, "invalid separator <%c> in <%s>\n",
*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
av = s+1;
}
if (i > 0) {
if (i+1 > F_LEN_MASK)
errx(EX_DATAERR, "too many port range\n");
cmd->o.len |= i+1; /* leave F_NOT and F_OR untouched */
}
return i;
}
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);
if (val <= 0)
errx(EX_DATAERR, "unknown ICMP unreachable code ``%s''", str);
*codep = val;
return;
}
static void
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
print_reject_code(u_int16_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
{
char *s = match_value(icmpcodes, code);
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.
*/
static int
contigmask(u_char *p, int len)
{
int i, n;
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
print_flags(char *name, ipfw_insn *cmd, struct _s_x *list)
{
char *comma="";
int i;
u_char set = cmd->arg1 & 0xff;
u_char clear = (cmd->arg1 >> 8) & 0xff;
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
print_ip(ipfw_insn_ip *cmd)
{
struct hostent *he = NULL;
int mb;
printf("%s ", cmd->o.len & F_NOT ? " not": "");
if (cmd->o.opcode == O_IP_SRC_ME || cmd->o.opcode == O_IP_DST_ME) {
printf("me");
return;
}
if (cmd->o.opcode == O_IP_SRC_SET || cmd->o.opcode == O_IP_DST_SET) {
u_int32_t x, *d;
int i;
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),
contigmask((u_char *)&x, 32));
x = cmd->addr.s_addr = htonl(cmd->addr.s_addr);
x &= 0xff; /* base */
d = (u_int32_t *)&(cmd->mask);
for (i=0; i < cmd->o.arg1; i++)
if (d[ i/32] & (1<<(i & 31))) {
printf("%c%d", comma, i+x);
comma = ',';
}
printf("}");
return;
}
if (cmd->o.opcode == O_IP_SRC || cmd->o.opcode == O_IP_DST)
mb = 32;
else
mb = contigmask((u_char *)&(cmd->mask.s_addr), 32);
if (mb == 32 && do_resolv)
he = gethostbyaddr((char *)&(cmd->addr.s_addr),
sizeof(u_long), AF_INET);
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 */
printf("%s", inet_ntoa(cmd->addr));
if (mb < 0)
printf(":%s", inet_ntoa(cmd->mask));
else if (mb < 32)
printf("/%d", mb);
}
}
/*
* prints a MAC address/mask pair
*/
static void
print_mac(u_char *addr, u_char *mask)
{
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)
{
u_int8_t type;
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.
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
*
* Special case: if we have provided a MAC header, and no IP specs,
* just leave it alone.
* Also, if we have providea a MAC header and no IP protocol, print it
* as "all" instead of "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
*/
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
#define HAVE_MAC 0x0008
#define HAVE_MACTYPE 0x0010
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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_IP (HAVE_PROTO | HAVE_SRCIP | HAVE_DSTIP)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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
show_prerequisites(int *flags, int want)
{
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 ( (*flags & (HAVE_MAC | HAVE_MACTYPE)) == HAVE_MAC) {
printf(" any"); /* MAC type */
*flags |= HAVE_MACTYPE;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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_PROTO) && (want & HAVE_PROTO))
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( (*flags & HAVE_MAC) ? " all" : " 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
if ( !(*flags & HAVE_SRCIP) && (want & HAVE_SRCIP))
printf(" from any");
if ( !(*flags & HAVE_DSTIP) && (want & HAVE_DSTIP))
printf(" to any");
*flags |= want;
}
static void
show_ipfw(struct ip_fw *rule)
{
int l;
ipfw_insn *cmd;
int proto = 0; /* default */
int flags = 0; /* prerequisites */
ipfw_insn_log *logptr = NULL; /* set if we find an O_LOG */
int or_block = 0; /* we are in an or block */
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
u_int32_t set_disable = (u_int32_t)(rule->next_rule);
if (set_disable & (1 << rule->set)) { /* disabled */
if (!show_sets)
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);
if (do_acct)
printf("%10qu %10qu ", rule->pcnt, rule->bcnt);
if (do_time) {
if (rule->timestamp) {
char timestr[30];
time_t t = _long_to_time(rule->timestamp);
strcpy(timestr, ctime(&t));
*strchr(timestr, '\n') = '\0';
printf("%s ", timestr);
} else {
printf(" ");
}
}
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 (show_sets)
printf("set %d ", rule->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
/*
* 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");
/* avoid printing anything else */
flags = HAVE_PROTO|HAVE_SRCIP|HAVE_DSTIP;
break;
case O_PROB:
{
ipfw_insn_u32 *p = (ipfw_insn_u32 *)cmd;
double d = 1.0 * p->d[0];
d = 1 - (d / 0x7fffffff);
printf("prob %f ", d);
}
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;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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_SKIPTO:
printf("skipto %u", cmd->arg1);
break;
case O_PIPE:
printf("pipe %u", cmd->arg1);
break;
case O_QUEUE:
printf("queue %u", cmd->arg1);
break;
case O_DIVERT:
printf("divert %u", cmd->arg1);
break;
case O_TEE:
printf("tee %u", cmd->arg1);
break;
case O_FORWARD_IP:
{
ipfw_insn_sa *s = (ipfw_insn_sa *)cmd;
printf("fwd %s", inet_ntoa(s->sa.sin_addr));
if (s->sa.sin_port)
printf(",%d", ntohs(s->sa.sin_port));
}
break;
case O_LOG: /* O_LOG is printed last */
logptr = (ipfw_insn_log *)cmd;
break;
default:
printf("** unrecognized action %d len %d",
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
}
/*
* then print the body
*/
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
switch(cmd->opcode) {
case O_PROBE_STATE:
break; /* no need to print anything here */
case O_MACADDR2: {
ipfw_insn_mac *m = (ipfw_insn_mac *)cmd;
if ( (flags & HAVE_MAC) == 0)
printf(" MAC");
flags |= HAVE_MAC;
if (cmd->len & F_NOT)
printf(" not");
print_mac( m->addr, m->mask);
print_mac( m->addr + 6, m->mask + 6);
}
break;
case O_MAC_TYPE:
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 ( (flags & HAVE_MAC) == 0)
printf(" MAC");
flags |= (HAVE_MAC | HAVE_MACTYPE);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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_newports((ipfw_insn_u16 *)cmd, IPPROTO_ETHERTYPE);
break;
case O_IP_SRC:
case O_IP_SRC_MASK:
case O_IP_SRC_ME:
case O_IP_SRC_SET:
show_prerequisites(&flags, HAVE_PROTO);
if (!(flags & HAVE_SRCIP))
printf(" from");
if ((cmd->len & F_OR) && !or_block)
printf(" {");
print_ip((ipfw_insn_ip *)cmd);
flags |= HAVE_SRCIP;
break;
case O_IP_DST:
case O_IP_DST_MASK:
case O_IP_DST_ME:
case O_IP_DST_SET:
show_prerequisites(&flags, HAVE_PROTO|HAVE_SRCIP);
if (!(flags & HAVE_DSTIP))
printf(" to");
if ((cmd->len & F_OR) && !or_block)
printf(" {");
print_ip((ipfw_insn_ip *)cmd);
flags |= HAVE_DSTIP;
break;
case O_IP_DSTPORT:
show_prerequisites(&flags,
HAVE_PROTO|HAVE_SRCIP|HAVE_DSTIP);
case O_IP_SRCPORT:
show_prerequisites(&flags, HAVE_PROTO|HAVE_SRCIP);
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(" {");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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_newports((ipfw_insn_u16 *)cmd, proto);
break;
case O_PROTO: {
struct protoent *pe;
if ((cmd->len & F_OR) && !or_block)
printf(" {");
if (cmd->len & F_NOT)
printf(" not");
proto = cmd->arg1;
pe = getprotobynumber(cmd->arg1);
if (pe)
printf(" %s", pe->p_name);
else
printf(" %u", cmd->arg1);
}
flags |= HAVE_PROTO;
break;
default: /*options ... */
show_prerequisites(&flags,
HAVE_PROTO|HAVE_SRCIP|HAVE_DSTIP);
if ((cmd->len & F_OR) && !or_block)
printf(" {");
if (cmd->len & F_NOT && cmd->opcode != O_IN)
printf(" not");
switch(cmd->opcode) {
case O_FRAG:
printf(" frag");
break;
case O_IN:
printf(cmd->len & F_NOT ? " out" : " in");
break;
case O_LAYER2:
printf(" layer2");
break;
case O_XMIT:
case O_RECV:
case O_VIA: {
char *s;
ipfw_insn_if *cmdif = (ipfw_insn_if *)cmd;
if (cmd->opcode == O_XMIT)
s = "xmit";
else if (cmd->opcode == O_RECV)
s = "recv";
else if (cmd->opcode == O_VIA)
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));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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 (cmdif->p.unit == -1)
printf(" %s %s*", s, cmdif->name);
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(" %s %s%d", s, cmdif->name,
cmdif->p.unit);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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_IPID:
printf(" ipid %u", cmd->arg1 );
break;
case O_IPTTL:
printf(" ipttl %u", cmd->arg1 );
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:
printf(" iplen %u", cmd->arg1 );
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;
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;
case O_KEEP_STATE:
printf(" keep-state");
break;
case O_LIMIT:
{
struct _s_x *p = limit_masks;
ipfw_insn_limit *c = (ipfw_insn_limit *)cmd;
u_int8_t x = c->limit_mask;
char *comma = " ";
printf(" limit");
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 = ",";
}
printf(" %d", c->conn_limit);
}
break;
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;
}
}
show_prerequisites(&flags, HAVE_PROTO|HAVE_SRCIP|HAVE_DSTIP);
printf("\n");
}
static void
show_dyn_ipfw(ipfw_dyn_rule *d)
{
struct protoent *pe;
struct in_addr a;
if (!do_expired) {
if (!d->expire && !(d->dyn_type == O_LIMIT_PARENT))
return;
}
printf("%05d %10qu %10qu (%ds)",
(int)(d->rule), d->pcnt, d->bcnt, d->expire);
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 */
printf(" STATE");
break;
}
if ((pe = getprotobynumber(d->id.proto)) != NULL)
printf(" %s", pe->p_name);
else
printf(" proto %u", d->id.proto);
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);
printf("\n");
}
int
sort_q(const void *pa, const void *pb)
{
int rev = (do_sort < 0);
int field = rev ? -do_sort : do_sort;
long long res = 0;
const struct dn_flow_queue *a = pa;
const struct dn_flow_queue *b = pb;
switch (field) {
case 1: /* pkts */
res = a->len - b->len;
break;
case 2: /* bytes */
res = a->len_bytes - b->len_bytes;
break;
case 3: /* tot pkts */
res = a->tot_pkts - b->tot_pkts;
break;
case 4: /* tot bytes */
res = a->tot_bytes - b->tot_bytes;
break;
}
if (res < 0)
res = -1;
if (res > 0)
res = 1;
return (int)(rev ? res : -res);
}
static void
list_queues(struct dn_flow_set *fs, struct dn_flow_queue *q)
{
int l;
printf(" mask: 0x%02x 0x%08x/0x%04x -> 0x%08x/0x%04x\n",
fs->flow_mask.proto,
fs->flow_mask.src_ip, fs->flow_mask.src_port,
fs->flow_mask.dst_ip, fs->flow_mask.dst_port);
if (fs->rq_elements == 0)
return;
printf("BKT Prot ___Source IP/port____ "
"____Dest. IP/port____ Tot_pkt/bytes Pkt/Byte Drp\n");
if (do_sort != 0)
heapsort(q, fs->rq_elements, sizeof *q, sort_q);
for (l = 0; l < fs->rq_elements; l++) {
struct in_addr ina;
struct protoent *pe;
ina.s_addr = htonl(q[l].id.src_ip);
printf("%3d ", q[l].hash_slot);
pe = getprotobynumber(q[l].id.proto);
if (pe)
printf("%-4s ", pe->p_name);
else
printf("%4u ", q[l].id.proto);
printf("%15s/%-5d ",
inet_ntoa(ina), q[l].id.src_port);
ina.s_addr = htonl(q[l].id.dst_ip);
printf("%15s/%-5d ",
inet_ntoa(ina), q[l].id.dst_port);
printf("%4qu %8qu %2u %4u %3u\n",
q[l].tot_pkts, q[l].tot_bytes,
q[l].len, q[l].len_bytes, q[l].drops);
if (verbose)
printf(" S %20qd F %20qd\n",
q[l].S, q[l].F);
}
}
static void
print_flowset_parms(struct dn_flow_set *fs, char *prefix)
{
int l;
char qs[30];
char plr[30];
char red[90]; /* Display RED parameters */
l = fs->qsize;
if (fs->flags_fs & DN_QSIZE_IS_BYTES) {
if (l >= 8192)
sprintf(qs, "%d KB", l / 1024);
else
sprintf(qs, "%d B", l);
} else
sprintf(qs, "%3d sl.", l);
if (fs->plr)
sprintf(plr, "plr %f", 1.0 * fs->plr / (double)(0x7fffffff));
else
plr[0] = '\0';
if (fs->flags_fs & DN_IS_RED) /* RED parameters */
sprintf(red,
"\n\t %cRED w_q %f min_th %d max_th %d max_p %f",
(fs->flags_fs & DN_IS_GENTLE_RED) ? 'G' : ' ',
1.0 * fs->w_q / (double)(1 << SCALE_RED),
SCALE_VAL(fs->min_th),
SCALE_VAL(fs->max_th),
1.0 * fs->max_p / (double)(1 << SCALE_RED));
else
sprintf(red, "droptail");
printf("%s %s%s %d queues (%d buckets) %s\n",
prefix, qs, plr, fs->rq_elements, fs->rq_size, red);
}
static void
list_pipes(void *data, int nbytes, int ac, char *av[])
{
u_long rulenum;
void *next = data;
struct dn_pipe *p = (struct dn_pipe *) data;
struct dn_flow_set *fs;
struct dn_flow_queue *q;
int l;
if (ac > 0)
rulenum = strtoul(*av++, NULL, 10);
else
rulenum = 0;
for (; nbytes >= sizeof *p; p = (struct dn_pipe *)next) {
double b = p->bandwidth;
char buf[30];
char prefix[80];
if (p->next != (struct dn_pipe *)DN_IS_PIPE)
break; /* done with pipes, now queues */
/*
* compute length, as pipe have variable size
*/
l = sizeof(*p) + p->fs.rq_elements * sizeof(*q);
next = (void *)p + l;
nbytes -= l;
if (rulenum != 0 && rulenum != p->pipe_nr)
continue;
/*
* Print rate (or clocking interface)
*/
if (p->if_name[0] != '\0')
sprintf(buf, "%s", p->if_name);
else if (b == 0)
sprintf(buf, "unlimited");
else if (b >= 1000000)
sprintf(buf, "%7.3f Mbit/s", b/1000000);
else if (b >= 1000)
sprintf(buf, "%7.3f Kbit/s", b/1000);
else
sprintf(buf, "%7.3f bit/s ", b);
sprintf(prefix, "%05d: %s %4d ms ",
p->pipe_nr, buf, p->delay);
print_flowset_parms(&(p->fs), prefix);
if (verbose)
printf(" V %20qd\n", p->V >> MY_M);
q = (struct dn_flow_queue *)(p+1);
list_queues(&(p->fs), q);
}
for (fs = next; nbytes >= sizeof *fs; fs = next) {
char prefix[80];
if (fs->next != (struct dn_flow_set *)DN_IS_QUEUE)
break;
l = sizeof(*fs) + fs->rq_elements * sizeof(*q);
next = (void *)fs + l;
nbytes -= l;
q = (struct dn_flow_queue *)(fs+1);
sprintf(prefix, "q%05d: weight %d pipe %d ",
fs->fs_nr, fs->weight, fs->parent_nr);
print_flowset_parms(fs, prefix);
list_queues(fs, q);
}
}
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
*/
static void
sets_handler(int ac, char *av[])
{
u_int32_t set_disable, masks[2];
int i, nbytes;
u_int16_t rulenum;
u_int8_t cmd, new_set;
ac--;
av++;
if (!ac)
errx(EX_USAGE, "set needs command");
if (!strncmp(*av, "show", strlen(*av)) ) {
void *data;
char *msg;
nbytes = sizeof(struct ip_fw);
if ((data = malloc(nbytes)) == NULL)
err(EX_OSERR, "malloc");
if (getsockopt(s, IPPROTO_IP, IP_FW_GET, data, &nbytes) < 0)
err(EX_OSERR, "getsockopt(IP_FW_GET)");
set_disable = (u_int32_t)(((struct ip_fw *)data)->next_rule);
for (i = 0, msg = "disable" ; i < 31; i++)
if ( (set_disable & (1<<i))) {
printf("%s %d", msg, i);
msg = "";
}
msg = (set_disable) ? " enable" : "enable";
for (i = 0; i < 31; i++)
if ( !(set_disable & (1<<i))) {
printf("%s %d", msg, i);
msg = "";
}
printf("\n");
} else if (!strncmp(*av, "swap", strlen(*av))) {
ac--; av++;
if (ac != 2)
errx(EX_USAGE, "set swap needs 2 set numbers\n");
rulenum = atoi(av[0]);
new_set = atoi(av[1]);
if (!isdigit(*(av[0])) || rulenum > 30)
errx(EX_DATAERR, "invalid set number %s\n", av[0]);
if (!isdigit(*(av[1])) || new_set > 30)
errx(EX_DATAERR, "invalid set number %s\n", av[1]);
masks[0] = (4 << 24) | (new_set << 16) | (rulenum);
i = setsockopt(s, IPPROTO_IP, IP_FW_DEL,
masks, sizeof(u_int32_t));
} else if (!strncmp(*av, "move", strlen(*av))) {
ac--; av++;
if (ac && !strncmp(*av, "rule", strlen(*av))) {
cmd = 2;
ac--; av++;
} else
cmd = 3;
if (ac != 3 || strncmp(av[1], "to", strlen(*av)))
errx(EX_USAGE, "syntax: set move [rule] X to Y\n");
rulenum = atoi(av[0]);
new_set = atoi(av[2]);
if (!isdigit(*(av[0])) || (cmd == 3 && rulenum > 30) ||
(cmd == 2 && rulenum == 65535) )
errx(EX_DATAERR, "invalid source number %s\n", av[0]);
if (!isdigit(*(av[2])) || new_set > 30)
errx(EX_DATAERR, "invalid dest. set %s\n", av[1]);
masks[0] = (cmd << 24) | (new_set << 16) | (rulenum);
i = setsockopt(s, IPPROTO_IP, IP_FW_DEL,
masks, sizeof(u_int32_t));
} else if (!strncmp(*av, "disable", strlen(*av)) ||
!strncmp(*av, "enable", strlen(*av)) ) {
int which = !strncmp(*av, "enable", strlen(*av)) ? 1 : 0;
ac--; av++;
masks[0] = masks[1] = 0;
while (ac) {
if (isdigit(**av)) {
i = atoi(*av);
if (i < 0 || i > 30)
errx(EX_DATAERR,
"invalid set number %d\n", i);
masks[which] |= (1<<i);
} else if (!strncmp(*av, "disable", strlen(*av)))
which = 0;
else if (!strncmp(*av, "enable", strlen(*av)))
which = 1;
else
errx(EX_DATAERR,
"invalid set command %s\n", *av);
av++; ac--;
}
if ( (masks[0] & masks[1]) != 0 )
errx(EX_DATAERR,
"cannot enable and disable the same set\n");
i = setsockopt(s, IPPROTO_IP, IP_FW_DEL, masks, sizeof(masks));
if (i)
warn("set enable/disable: setsockopt(IP_FW_DEL)");
} else
errx(EX_USAGE, "invalid set command %s\n", *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
static void
list(int ac, char *av[])
{
struct ip_fw *r;
ipfw_dyn_rule *dynrules, *d;
void *lim, *data = NULL;
int n, nbytes, nstat, ndyn;
int exitval = EX_OK;
int lac;
char **lav;
u_long rnum;
char *endptr;
int seen = 0;
const int ocmd = do_pipe ? IP_DUMMYNET_GET : IP_FW_GET;
int nalloc = 1024; /* start somewhere... */
ac--;
av++;
/* get rules or pipes from kernel, resizing array as necessary */
nbytes = nalloc;
while (nbytes >= nalloc) {
nalloc = nalloc * 2 + 200;
nbytes = nalloc;
if ((data = realloc(data, nbytes)) == NULL)
err(EX_OSERR, "realloc");
if (getsockopt(s, IPPROTO_IP, ocmd, data, &nbytes) < 0)
err(EX_OSERR, "getsockopt(IP_%s_GET)",
do_pipe ? "DUMMYNET" : "FW");
}
if (do_pipe) {
list_pipes(data, nbytes, ac, av);
goto done;
}
/*
* Count static rules. They have variable size so we
* need to scan the list to count them.
*/
for (nstat = 1, r = data, lim = data + nbytes;
r->rulenum < 65535 && (void *)r < lim;
++nstat, r = (void *)r + RULESIZE(r) )
; /* nothing */
/*
* Count dynamic rules. This is easier as they have
* fixed size.
*/
r = (void *)r + RULESIZE(r);
dynrules = (ipfw_dyn_rule *)r ;
n = (void *)r - data;
ndyn = (nbytes - n) / sizeof *dynrules;
/* if no rule numbers were specified, list all rules */
if (ac == 0) {
for (n = 0, r = data; n < nstat;
n++, r = (void *)r + RULESIZE(r) )
show_ipfw(r);
if (do_dynamic && ndyn) {
printf("## Dynamic rules (%d):\n", ndyn);
for (n = 0, d = dynrules; n < ndyn; n++, d++)
show_dyn_ipfw(d);
}
goto done;
}
/* display specific rules requested on command line */
for (lac = ac, lav = av; lac != 0; lac--) {
/* convert command line rule # */
rnum = strtoul(*lav++, &endptr, 10);
if (*endptr) {
exitval = EX_USAGE;
warnx("invalid rule number: %s", *(lav - 1));
continue;
}
for (n = seen = 0, r = data; n < nstat;
n++, r = (void *)r + RULESIZE(r) ) {
if (r->rulenum > rnum)
break;
if (r->rulenum == rnum) {
show_ipfw(r);
seen = 1;
}
}
if (!seen) {
/* give precedence to other error(s) */
if (exitval == EX_OK)
exitval = EX_UNAVAILABLE;
warnx("rule %lu does not exist", rnum);
}
}
if (do_dynamic && ndyn) {
printf("## Dynamic rules:\n");
for (lac = ac, lav = av; lac != 0; lac--) {
rnum = strtoul(*lav++, &endptr, 10);
if (*endptr)
/* already warned */
continue;
for (n = 0, d = dynrules; n < ndyn; n++, d++) {
if ((int)(d->rule) > rnum)
break;
if ((int)(d->rule) == rnum)
show_dyn_ipfw(d);
}
}
}
ac = 0;
done:
free(data);
if (exitval != EX_OK)
exit(exitval);
}
static void
show_usage(void)
{
fprintf(stderr, "usage: ipfw [options]\n"
" add [number] rule\n"
" pipe number config [pipeconfig]\n"
" queue number config [queueconfig]\n"
" [pipe] flush\n"
" [pipe] delete number ...\n"
" [pipe] {list|show} [number ...]\n"
" {zero|resetlog} [number ...]\n"
"do \"ipfw -h\" or see ipfw manpage for details\n"
);
exit(EX_USAGE);
}
static void
help(void)
{
fprintf(stderr, "ipfw syntax summary:\n"
"ipfw add [N] [prob {0..1}] ACTION [log [logamount N]] ADDR OPTIONS\n"
"ipfw {pipe|queue} N config BODY\n"
"ipfw [pipe] {zero|delete|show} [N{,N}]\n"
"\n"
"RULE: [1..] [PROB] BODY\n"
"RULENUM: INTEGER(1..65534)\n"
"PROB: prob REAL(0..1)\n"
"BODY: check-state [LOG] (no body) |\n"
" ACTION [LOG] MATCH_ADDR [OPTION_LIST]\n"
"ACTION: check-state | allow | count | deny | reject | skipto N |\n"
" {divert|tee} PORT | forward ADDR | pipe N | queue N\n"
"ADDR: [ MAC dst src ether_type ] \n"
" [ from IPLIST [ PORT ] to IPLIST [ PORTLIST ] ]\n"
"IPLIST: IPADDR | ( IPADDR or ... or IPADDR )\n"
"IPADDR: [not] { any | me | ip | ip/bits | ip:mask | ip/bits{x,y,z} }\n"
"OPTION_LIST: OPTION [,OPTION_LIST]\n"
);
exit(0);
}
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:
* any matches any IP. Actually returns an empty instruction.
* 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
*/
static void
fill_ip(ipfw_insn_ip *cmd, char *av)
{
char *p = 0, md = 0;
u_int32_t i;
cmd->o.len &= ~F_LEN_MASK; /* zero len */
if (!strncmp(av, "any", strlen(av)))
return;
if (!strncmp(av, "me", strlen(av))) {
cmd->o.len |= F_INSN_SIZE(ipfw_insn);
return;
}
p = strchr(av, '/');
if (!p)
p = strchr(av, ':');
if (p) {
md = *p;
*p++ = '\0';
}
if (lookup_host(av, &cmd->addr) != 0)
errx(EX_NOHOST, "hostname ``%s'' unknown", av);
switch (md) {
case ':':
if (!inet_aton(p, &cmd->mask))
errx(EX_DATAERR, "bad netmask ``%s''", p);
break;
case '/':
i = atoi(p);
if (i == 0)
cmd->mask.s_addr = htonl(0);
else if (i > 32)
errx(EX_DATAERR, "bad width ``%s''", p);
else
cmd->mask.s_addr = htonl(~0 << (32 - i));
break;
default:
cmd->mask.s_addr = htonl(~0);
break;
}
cmd->addr.s_addr &= cmd->mask.s_addr;
/*
* now look if 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 u_int32_t with bits set.
*/
if (p)
p = strchr(p, '{');
if (p) { /* fetch addresses */
u_int32_t *d;
int low, high;
int i = contigmask((u_char *)&(cmd->mask), 32);
if (i < 24 || i > 31) {
fprintf(stderr, "invalid set with mask %d\n",
i);
exit(0);
}
cmd->o.arg1 = 1<<(32-i);
cmd->addr.s_addr = ntohl(cmd->addr.s_addr);
d = (u_int32_t *)&cmd->mask;
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++)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
d[i] = 0; /* clear masks */
av = p+1;
low = cmd->addr.s_addr & 0xff;
high = low + cmd->o.arg1 - 1;
while (isdigit(*av)) {
char *s;
u_int16_t a = strtol(av, &s, 0);
if (s == av) /* no parameter */
break;
if (a < low || a > high) {
fprintf(stderr, "addr %d out of range [%d-%d]\n",
a, low, high);
exit(0);
}
a -= low;
d[ a/32] |= 1<<(a & 31);
if (*s != ',')
break;
av = s+1;
}
return;
}
if (cmd->mask.s_addr == 0) { /* any */
if (cmd->o.len & F_NOT)
errx(EX_DATAERR, "not any never matches");
else /* useless, nuke it */
return;
} else if (cmd->mask.s_addr == IP_MASK_ALL) /* one IP */
cmd->o.len |= F_INSN_SIZE(ipfw_insn_u32);
else /* addr/mask */
cmd->o.len |= F_INSN_SIZE(ipfw_insn_ip);
}
/*
* 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)
{
u_int8_t set=0, clear=0;
while (p && *p) {
char *q; /* points to the separator */
int val;
u_int8_t *which; /* mask we are working on */
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);
*which |= (u_int8_t)val;
p = q;
}
cmd->opcode = opcode;
cmd->len = (cmd->len & (F_NOT | F_OR)) | 1;
cmd->arg1 = (set & 0xff) | ( (clear & 0xff) << 8);
}
static void
delete(int ac, 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
u_int32_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
struct dn_pipe pipe;
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
memset(&pipe, 0, sizeof pipe);
av++; ac--;
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 (ac > 0 && !strncmp(*av, "set", strlen(*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
do_set = 1; /* delete 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
ac--; 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 */
while (ac && isdigit(**av)) {
i = atoi(*av); av++; ac--;
if (do_pipe) {
if (do_pipe == 1)
pipe.pipe_nr = i;
else
pipe.fs.fs_nr = i;
i = setsockopt(s, IPPROTO_IP, IP_DUMMYNET_DEL,
&pipe, sizeof pipe);
if (i) {
exitval = 1;
warn("rule %u: setsockopt(IP_DUMMYNET_DEL)",
do_pipe == 1 ? pipe.pipe_nr :
pipe.fs.fs_nr);
}
} 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);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
i = setsockopt(s, IPPROTO_IP, IP_FW_DEL, &rulenum,
sizeof rulenum);
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.
* A '*' following the name means any unit.
*/
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 */
if (!strcmp(arg, "any"))
cmd->o.len = 0; /* effectively ignore this command */
else if (!isdigit(*arg)) {
char *q;
strncpy(cmd->name, arg, sizeof(cmd->name));
cmd->name[sizeof(cmd->name) - 1] = '\0';
/* find first digit or wildcard */
for (q = cmd->name; *q && !isdigit(*q) && *q != '*'; q++)
continue;
cmd->p.unit = (*q == '*') ? -1 : atoi(q);
*q = '\0';
} else if (!inet_aton(arg, &cmd->p.ip))
errx(EX_DATAERR, "bad ip address ``%s''", arg);
}
/*
* the following macro returns an error message if we run out of
* arguments.
*/
#define NEED1(msg) {if (!ac) errx(EX_USAGE, msg);}
static void
config_pipe(int ac, char **av)
{
struct dn_pipe pipe;
int i;
char *end;
u_int32_t a;
void *par = NULL;
memset(&pipe, 0, sizeof pipe);
av++; ac--;
/* Pipe number */
if (ac && isdigit(**av)) {
i = atoi(*av); av++; ac--;
if (do_pipe == 1)
pipe.pipe_nr = i;
else
pipe.fs.fs_nr = i;
}
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
while (ac > 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
double d;
int tok = match_token(dummynet_params, *av);
ac--; av++;
switch(tok) {
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
case TOK_NOERROR:
pipe.fs.flags_fs |= DN_NOERROR;
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_PLR:
NEED1("plr needs argument 0..1\n");
d = strtod(av[0], NULL);
if (d > 1)
d = 1;
else if (d < 0)
d = 0;
pipe.fs.plr = (int)(d*0x7fffffff);
ac--; av++;
break;
case TOK_QUEUE:
NEED1("queue needs queue size\n");
end = NULL;
pipe.fs.qsize = strtoul(av[0], &end, 0);
if (*end == 'K' || *end == 'k') {
pipe.fs.flags_fs |= DN_QSIZE_IS_BYTES;
pipe.fs.qsize *= 1024;
} else if (*end == 'B' || !strncmp(end, "by", 2)) {
pipe.fs.flags_fs |= DN_QSIZE_IS_BYTES;
}
ac--; av++;
break;
case TOK_BUCKETS:
NEED1("buckets needs argument\n");
pipe.fs.rq_size = strtoul(av[0], NULL, 0);
ac--; av++;
break;
case TOK_MASK:
NEED1("mask needs mask specifier\n");
/*
* per-flow queue, mask is dst_ip, dst_port,
* src_ip, src_port, proto measured in bits
*/
par = NULL;
pipe.fs.flow_mask.dst_ip = 0;
pipe.fs.flow_mask.src_ip = 0;
pipe.fs.flow_mask.dst_port = 0;
pipe.fs.flow_mask.src_port = 0;
pipe.fs.flow_mask.proto = 0;
end = NULL;
while (ac >= 1) {
u_int32_t *p32 = NULL;
u_int16_t *p16 = NULL;
tok = match_token(dummynet_params, *av);
ac--; av++;
switch(tok) {
case TOK_ALL:
/*
* special case, all bits significant
*/
pipe.fs.flow_mask.dst_ip = ~0;
pipe.fs.flow_mask.src_ip = ~0;
pipe.fs.flow_mask.dst_port = ~0;
pipe.fs.flow_mask.src_port = ~0;
pipe.fs.flow_mask.proto = ~0;
pipe.fs.flags_fs |= DN_HAVE_FLOW_MASK;
goto end_mask;
case TOK_DSTIP:
p32 = &pipe.fs.flow_mask.dst_ip;
break;
case TOK_SRCIP:
p32 = &pipe.fs.flow_mask.src_ip;
break;
case TOK_DSTPORT:
p16 = &pipe.fs.flow_mask.dst_port;
break;
case TOK_SRCPORT:
p16 = &pipe.fs.flow_mask.src_port;
break;
case TOK_PROTO:
break;
default:
ac++; av--; /* backtrack */
goto end_mask;
}
if (ac < 1)
errx(EX_USAGE, "mask: value missing");
if (*av[0] == '/') {
a = strtoul(av[0]+1, &end, 0);
a = (a == 32) ? ~0 : (1 << a) - 1;
} 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
a = strtoul(av[0], &end, 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 (p32 != NULL)
*p32 = a;
else if (p16 != NULL) {
if (a > 65535)
errx(EX_DATAERR,
"mask: must be 16 bit");
*p16 = (u_int16_t)a;
} else {
if (a > 255)
errx(EX_DATAERR,
"mask: must be 8 bit");
pipe.fs.flow_mask.proto = (u_int8_t)a;
}
if (a != 0)
pipe.fs.flags_fs |= DN_HAVE_FLOW_MASK;
ac--; av++;
} /* end while, config masks */
end_mask:
break;
case TOK_RED:
case TOK_GRED:
NEED1("red/gred needs w_q/min_th/max_th/max_p\n");
pipe.fs.flags_fs |= DN_IS_RED;
if (tok == TOK_GRED)
pipe.fs.flags_fs |= DN_IS_GENTLE_RED;
/*
* the format for parameters is w_q/min_th/max_th/max_p
*/
if ((end = strsep(&av[0], "/"))) {
double w_q = strtod(end, NULL);
if (w_q > 1 || w_q <= 0)
errx(EX_DATAERR, "0 < w_q <= 1");
pipe.fs.w_q = (int) (w_q * (1 << SCALE_RED));
}
if ((end = strsep(&av[0], "/"))) {
pipe.fs.min_th = strtoul(end, &end, 0);
if (*end == 'K' || *end == 'k')
pipe.fs.min_th *= 1024;
}
if ((end = strsep(&av[0], "/"))) {
pipe.fs.max_th = strtoul(end, &end, 0);
if (*end == 'K' || *end == 'k')
pipe.fs.max_th *= 1024;
}
if ((end = strsep(&av[0], "/"))) {
double max_p = strtod(end, NULL);
if (max_p > 1 || max_p <= 0)
errx(EX_DATAERR, "0 < max_p <= 1");
pipe.fs.max_p = (int)(max_p * (1 << SCALE_RED));
}
ac--; av++;
break;
case TOK_DROPTAIL:
pipe.fs.flags_fs &= ~(DN_IS_RED|DN_IS_GENTLE_RED);
break;
case TOK_BW:
NEED1("bw needs bandwidth or interface\n");
if (do_pipe != 1)
errx(EX_DATAERR, "bandwidth only valid for pipes");
/*
* set clocking interface or bandwidth value
*/
if (av[0][0] >= 'a' && av[0][0] <= 'z') {
int l = sizeof(pipe.if_name)-1;
/* interface name */
strncpy(pipe.if_name, av[0], l);
pipe.if_name[l] = '\0';
pipe.bandwidth = 0;
} else {
pipe.if_name[0] = '\0';
pipe.bandwidth = strtoul(av[0], &end, 0);
if (*end == 'K' || *end == 'k') {
end++;
pipe.bandwidth *= 1000;
} else if (*end == 'M') {
end++;
pipe.bandwidth *= 1000000;
}
if (*end == 'B' || !strncmp(end, "by", 2))
pipe.bandwidth *= 8;
if (pipe.bandwidth < 0)
errx(EX_DATAERR, "bandwidth too large");
}
ac--; av++;
break;
case TOK_DELAY:
if (do_pipe != 1)
errx(EX_DATAERR, "delay only valid for pipes");
NEED1("delay needs argument 0..10000ms\n");
pipe.delay = strtoul(av[0], NULL, 0);
ac--; av++;
break;
case TOK_WEIGHT:
if (do_pipe == 1)
errx(EX_DATAERR,"weight only valid for queues");
NEED1("weight needs argument 0..100\n");
pipe.fs.weight = strtoul(av[0], &end, 0);
ac--; av++;
break;
case TOK_PIPE:
if (do_pipe == 1)
errx(EX_DATAERR,"pipe only valid for queues");
NEED1("pipe needs pipe_number\n");
pipe.fs.parent_nr = strtoul(av[0], &end, 0);
ac--; av++;
break;
default:
errx(EX_DATAERR, "unrecognised option ``%s''", *av);
}
}
if (do_pipe == 1) {
if (pipe.pipe_nr == 0)
errx(EX_DATAERR, "pipe_nr must be > 0");
if (pipe.delay > 10000)
errx(EX_DATAERR, "delay must be < 10000");
} else { /* do_pipe == 2, queue */
if (pipe.fs.parent_nr == 0)
errx(EX_DATAERR, "pipe must be > 0");
if (pipe.fs.weight >100)
errx(EX_DATAERR, "weight must be <= 100");
}
if (pipe.fs.flags_fs & DN_QSIZE_IS_BYTES) {
if (pipe.fs.qsize > 1024*1024)
errx(EX_DATAERR, "queue size must be < 1MB");
} else {
if (pipe.fs.qsize > 100)
errx(EX_DATAERR, "2 <= queue size <= 100");
}
if (pipe.fs.flags_fs & DN_IS_RED) {
size_t len;
int lookup_depth, avg_pkt_size;
double s, idle, weight, w_q;
struct clockinfo clock;
int t;
if (pipe.fs.min_th >= pipe.fs.max_th)
errx(EX_DATAERR, "min_th %d must be < than max_th %d",
pipe.fs.min_th, pipe.fs.max_th);
if (pipe.fs.max_th == 0)
errx(EX_DATAERR, "max_th must be > 0");
len = sizeof(int);
if (sysctlbyname("net.inet.ip.dummynet.red_lookup_depth",
&lookup_depth, &len, NULL, 0) == -1)
errx(1, "sysctlbyname(\"%s\")",
"net.inet.ip.dummynet.red_lookup_depth");
if (lookup_depth == 0)
errx(EX_DATAERR, "net.inet.ip.dummynet.red_lookup_depth"
" must be greater than zero");
len = sizeof(int);
if (sysctlbyname("net.inet.ip.dummynet.red_avg_pkt_size",
&avg_pkt_size, &len, NULL, 0) == -1)
errx(1, "sysctlbyname(\"%s\")",
"net.inet.ip.dummynet.red_avg_pkt_size");
if (avg_pkt_size == 0)
errx(EX_DATAERR,
"net.inet.ip.dummynet.red_avg_pkt_size must"
" be greater than zero");
len = sizeof(struct clockinfo);
if (sysctlbyname("kern.clockrate", &clock, &len, NULL, 0) == -1)
errx(1, "sysctlbyname(\"%s\")", "kern.clockrate");
/*
* Ticks needed for sending a medium-sized packet.
* Unfortunately, when we are configuring a WF2Q+ queue, we
* do not have bandwidth information, because that is stored
* in the parent pipe, and also we have multiple queues
* competing for it. So we set s=0, which is not very
* correct. But on the other hand, why do we want RED with
* WF2Q+ ?
*/
if (pipe.bandwidth==0) /* this is a WF2Q+ queue */
s = 0;
else
s = clock.hz * avg_pkt_size * 8 / pipe.bandwidth;
/*
* max idle time (in ticks) before avg queue size becomes 0.
* NOTA: (3/w_q) is approx the value x so that
* (1-w_q)^x < 10^-3.
*/
w_q = ((double)pipe.fs.w_q) / (1 << SCALE_RED);
idle = s * 3. / w_q;
pipe.fs.lookup_step = (int)idle / lookup_depth;
if (!pipe.fs.lookup_step)
pipe.fs.lookup_step = 1;
weight = 1 - w_q;
for (t = pipe.fs.lookup_step; t > 0; --t)
weight *= weight;
pipe.fs.lookup_weight = (int)(weight * (1 << SCALE_RED));
}
i = setsockopt(s, IPPROTO_IP, IP_DUMMYNET_CONFIGURE, &pipe,
sizeof pipe);
if (i)
err(1, "setsockopt(%s)", "IP_DUMMYNET_CONFIGURE");
}
static void
get_mac_addr_mask(char *p, u_char *addr, u_char *mask)
{
int i, l;
for (i=0; i<6; i++)
addr[i] = mask[i] = 0;
if (!strcmp(p, "any"))
return;
for (i=0; *p && i<6;i++, p++) {
addr[i] = strtol(p, &p, 16);
if (*p != ':') /* we start with the mask */
break;
}
if (*p == '/') { /* mask len */
l = strtol(p+1, &p, 0);
for (i=0; l>0; l -=8, i++)
mask[i] = (l >=8) ? 0xff : (~0) << (8-l);
} else if (*p == '&') { /* mask */
for (i=0, p++; *p && i<6;i++, p++) {
mask[i] = strtol(p, &p, 16);
if (*p != ':')
break;
}
} else if (*p == '\0') {
for (i=0; i<6; i++)
mask[i] = 0xff;
}
for (i=0; i<6; i++)
addr[i] &= mask[i];
}
/*
* 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;
}
/*
* A function to fill simple commands of size 1.
* Existing flags are preserved.
*/
static void
fill_cmd(ipfw_insn *cmd, enum ipfw_opcodes opcode, int flags, u_int16_t arg)
{
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 *
add_mac(ipfw_insn *cmd, int ac, char *av[])
{
ipfw_insn_mac *mac; /* also *src */
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 (ac <2)
errx(EX_DATAERR, "MAC dst src [type]");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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 */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
get_mac_addr_mask(av[1], &(mac->addr[6]), &(mac->mask[6])); /* src */
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 (ac>2 && strcmp(av[2], "any") != 0) { /* we have a non-null type */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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 += F_LEN(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
fill_newports((ipfw_insn_u16 *)cmd, av[2], 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;
}
return cmd;
}
/*
* 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.
*
* The syntax for a rule starts with the action, followed by an
* optional log action, and the various match patterns.
* In the assembled microcode, the first opcode must be a O_PROBE_STATE
* (generated if the rule includes a keep-state option), then the
* various match patterns, the "log" action, and the actual action.
*
*/
static void
add(int ac, char *av[])
{
/*
* 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[].
*/
static u_int32_t rulebuf[255], actbuf[255], cmdbuf[255];
ipfw_insn *src, *dst, *cmd, *action, *prev;
struct ip_fw *rule;
/*
* various flags used to record that we entered some fields.
*/
int have_mac = 0; /* set if we have a MAC address */
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 */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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 */
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;
av++; ac--;
/* [rule N] -- Rule number optional */
if (ac && isdigit(**av)) {
rule->rulenum = atoi(*av);
av++;
ac--;
}
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
/* [set N] -- set number (0..30), optional */
if (ac > 1 && !strncmp(*av, "set", strlen(*av))) {
int set = strtoul(av[1], NULL, 10);
if (set < 0 || set > 30)
errx(EX_DATAERR, "illegal set %s", av[1]);
rule->set = set;
av += 2; ac -= 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
/* [prob D] -- match probability, optional */
if (ac > 1 && !strncmp(*av, "prob", strlen(*av))) {
double d = strtod(av[1], NULL);
if (d <= 0 || d > 1)
errx(EX_DATAERR, "illegal match prob. %s", av[1]);
if (d != 1) { /* 1 means always match */
action->opcode = O_PROB;
action->len = 2;
*((int32_t *)(action+1)) =
(int32_t)((1 - d) * 0x7fffffff);
action += action->len;
}
av += 2; ac -= 2;
}
/* action -- mandatory */
NEED1("missing action");
i = match_token(rule_actions, *av);
ac--; av++;
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;
case TOK_UNREACH:
action->opcode = O_REJECT;
NEED1("missing reject code");
fill_reject_code(&action->arg1, *av);
ac--; 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_COUNT:
action->opcode = O_COUNT;
break;
case TOK_QUEUE:
case TOK_PIPE:
action->len = F_INSN_SIZE(ipfw_insn_pipe);
case TOK_SKIPTO:
if (i == TOK_QUEUE)
action->opcode = O_QUEUE;
else if (i == TOK_PIPE)
action->opcode = O_PIPE;
else if (i == TOK_SKIPTO)
action->opcode = O_SKIPTO;
NEED1("missing skipto/pipe/queue number");
action->arg1 = strtoul(*av, NULL, 10);
av++; ac--;
break;
case TOK_DIVERT:
case TOK_TEE:
action->opcode = (i == TOK_DIVERT) ? O_DIVERT : O_TEE;
NEED1("missing divert/tee port");
action->arg1 = strtoul(*av, NULL, 0);
if (action->arg1 == 0) {
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");
}
ac--; av++;
break;
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);
p->sa.sin_len = sizeof(struct sockaddr_in);
p->sa.sin_family = AF_INET;
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);
p->sa.sin_port = htons( (u_short)i );
}
lookup_host(*av, &(p->sa.sin_addr));
}
ac--; av++;
break;
default:
errx(EX_DATAERR, "invalid action %s\n", *av);
}
action = next_cmd(action);
/*
* [log [logamount N]] -- log, optional
*
* If exists, it goes first in the cmdbuf, but then it is
* skipped in the copy section to the end of the buffer.
*/
if (ac && !strncmp(*av, "log", strlen(*av))) {
ipfw_insn_log *c = (ipfw_insn_log *)cmd;
cmd->len = F_INSN_SIZE(ipfw_insn_log);
cmd->opcode = O_LOG;
av++; ac--;
if (ac && !strncmp(*av, "logamount", strlen(*av))) {
ac--; av++;
NEED1("logamount requires argument");
c->max_log = atoi(*av);
if (c->max_log < 0)
errx(EX_DATAERR, "logamount must be positive");
ac--; av++;
}
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) \
if (ac && (*av[0] == '(' || *av[0] == '{')) { \
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') { \
ac--; av++; \
} else \
(*av)++; \
} \
target: \
#define CLOSE_PAR \
if (open_par) { \
if (ac && ( \
!strncmp(*av, ")", strlen(*av)) || \
!strncmp(*av, "}", strlen(*av)) )) { \
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; \
ac--; av++; \
} else \
errx(EX_USAGE, "missing \")\"\n"); \
}
#define NOT_BLOCK \
if (ac && !strncmp(*av, "not", strlen(*av))) { \
if (cmd->len & F_NOT) \
errx(EX_USAGE, "double \"not\" not allowed\n"); \
cmd->len |= F_NOT; \
ac--; av++; \
}
#define OR_BLOCK(target) \
if (ac && !strncmp(*av, "or", strlen(*av))) { \
if (prev == NULL || open_par == 0) \
errx(EX_DATAERR, "invalid OR block"); \
prev->len |= F_OR; \
ac--; av++; \
goto target; \
} \
CLOSE_PAR;
/*
* protocol, mandatory
*/
OR_START(get_proto);
NOT_BLOCK;
NEED1("missing protocol");
{
struct protoent *pe;
if (!strncmp(*av, "all", strlen(*av)))
; /* same as "ip" */
else if (!strncmp(*av, "MAC", strlen(*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
cmd = add_mac(cmd, ac-1, av+1); /* exits in case of errors */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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 += 3;
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
ac -= 3;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
have_mac = 1;
} else if ((proto = atoi(*av)) > 0)
; /* all done! */
else if ((pe = getprotobyname(*av)) != NULL)
proto = pe->p_proto;
else
errx(EX_DATAERR, "invalid protocol ``%s''", *av);
av++; ac--;
if (proto != IPPROTO_IP)
fill_cmd(cmd, O_PROTO, 0, proto);
}
cmd = next_cmd(cmd);
OR_BLOCK(get_proto);
/*
* "from", mandatory (unless we have a MAC address)
*/
if (!ac || strncmp(*av, "from", strlen(*av))) {
if (have_mac) /* we do not need a "to" address */
goto read_to;
errx(EX_USAGE, "missing ``from''");
}
ac--; av++;
/*
* source IP, mandatory
*/
OR_START(source_ip);
NOT_BLOCK; /* optional "not" */
NEED1("missing source address");
/* source -- mandatory */
fill_ip((ipfw_insn_ip *)cmd, *av);
if (cmd->opcode == O_IP_DST_SET) /* set */
cmd->opcode = O_IP_SRC_SET;
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;
else if (F_LEN(cmd) == F_INSN_SIZE(ipfw_insn_ip)) /* addr/mask */
cmd->opcode = O_IP_SRC_MASK;
/* otherwise len will be zero and the command skipped */
ac--; av++;
prev = cmd; /* in case we need to backtrack */
cmd = next_cmd(cmd);
OR_BLOCK(source_ip);
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
OR_START(source_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
/*
* source ports, optional
*/
NOT_BLOCK; /* optional "not" */
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 (ac) {
if (!strncmp(*av, "any", strlen(*av))) {
ac--;
av++;
} else if (fill_newports((ipfw_insn_u16 *)cmd, *av, proto)) {
/* XXX todo: check that we have a protocol with ports */
cmd->opcode = O_IP_SRCPORT;
ac--;
av++;
cmd = next_cmd(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
}
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
OR_BLOCK(source_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
read_to:
/*
* "to", mandatory (unless we have a MAC address
*/
if (!ac || strncmp(*av, "to", strlen(*av))) {
if (have_mac)
goto read_options;
errx(EX_USAGE, "missing ``to''");
}
av++; ac--;
/*
* destination, mandatory
*/
OR_START(dest_ip);
NOT_BLOCK; /* optional "not" */
NEED1("missing dst address");
fill_ip((ipfw_insn_ip *)cmd, *av);
if (cmd->opcode == O_IP_DST_SET) /* set */
;
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;
else if (F_LEN(cmd) == F_INSN_SIZE(ipfw_insn_ip)) /* addr/mask */
cmd->opcode = O_IP_DST_MASK;
ac--;
av++;
prev = cmd;
cmd = next_cmd(cmd);
OR_BLOCK(dest_ip);
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
OR_START(dest_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
/*
* dest. ports, optional
*/
NOT_BLOCK; /* optional "not" */
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 (ac) {
if (!strncmp(*av, "any", strlen(*av))) {
ac--;
av++;
} else if (fill_newports((ipfw_insn_u16 *)cmd, *av, proto)) {
/* XXX todo: check that we have a protocol with ports */
cmd->opcode = O_IP_DSTPORT;
ac--;
av++;
cmd += F_LEN(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
}
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
OR_BLOCK(dest_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
read_options:
prev = NULL;
while (ac) {
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);
ac--; av++;
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;
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;
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]);
ac--; av++;
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);
av++; ac--;
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_IPTTL:
NEED1("ipttl requires TTL");
fill_cmd(cmd, O_IPTTL, 0, strtoul(*av, NULL, 0));
ac--; av++;
break;
case TOK_IPID:
NEED1("ipid requires length");
fill_cmd(cmd, O_IPID, 0, strtoul(*av, NULL, 0));
ac--; av++;
break;
case TOK_IPLEN:
NEED1("iplen requires length");
fill_cmd(cmd, O_IPLEN, 0, strtoul(*av, NULL, 0));
ac--; av++;
break;
case TOK_IPVER:
NEED1("ipver requires version");
fill_cmd(cmd, O_IPVER, 0, strtoul(*av, NULL, 0));
ac--; av++;
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);
ac--; av++;
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);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of 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++;
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);
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
ac--; av++;
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);
cmd32->d[0] = uid;
cmd->len = F_INSN_SIZE(ipfw_insn_u32);
ac--; av++;
}
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);
cmd32->d[0] = gid;
cmd->len = F_INSN_SIZE(ipfw_insn_u32);
ac--; av++;
}
break;
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;
case TOK_TCPOPTS:
NEED1("missing argument for tcpoptions");
fill_flags(cmd, O_TCPOPTS, f_tcpopts, *av);
ac--; av++;
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));
ac--; av++;
break;
case TOK_TCPWIN:
NEED1("tcpwin requires length");
fill_cmd(cmd, O_TCPWIN, 0,
htons(strtoul(*av, NULL, 0)));
ac--; av++;
break;
case TOK_TCPFLAGS:
NEED1("missing argument for tcpflags");
cmd->opcode = O_TCPFLAGS;
fill_flags(cmd, O_TCPFLAGS, f_tcpflags, *av);
ac--; av++;
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;
case TOK_LIMIT:
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, "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)
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");
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
NEED1("limit needs mask and # of connections");
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
{
ipfw_insn_limit *c = (ipfw_insn_limit *)cmd;
cmd->len = F_INSN_SIZE(ipfw_insn_limit);
cmd->opcode = O_LIMIT;
c->limit_mask = 0;
c->conn_limit = 0;
for (; ac >1 ;) {
int val;
val = match_token(limit_masks, *av);
if (val <= 0)
break;
c->limit_mask |= val;
ac--; av++;
}
c->conn_limit = atoi(*av);
if (c->conn_limit == 0)
errx(EX_USAGE, "limit: limit must be >0");
if (c->limit_mask == 0)
errx(EX_USAGE, "missing limit mask");
ac--; av++;
}
break;
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;
/*
* 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);
}
/*
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
* copy all commands but O_LOG, O_KEEP_STATE, 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
*/
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:
break;
default:
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
bcopy(src, dst, i * sizeof(u_int32_t));
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);
bcopy(have_state, dst, i * sizeof(u_int32_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
/*
* start action section
*/
rule->act_ofs = dst - rule->cmd;
/*
* put back O_LOG if necessary
*/
src = (ipfw_insn *)cmdbuf;
if ( src->opcode == O_LOG ) {
i = F_LEN(src);
bcopy(src, dst, i * sizeof(u_int32_t));
dst += i;
}
/*
* copy all other actions
*/
for (src = (ipfw_insn *)actbuf; src != action; src += i) {
i = F_LEN(src);
bcopy(src, dst, i * sizeof(u_int32_t));
dst += i;
}
rule->cmd_len = (u_int32_t *)dst - (u_int32_t *)(rule->cmd);
i = (void *)dst - (void *)rule;
if (getsockopt(s, IPPROTO_IP, IP_FW_ADD, rule, &i) == -1)
err(EX_UNAVAILABLE, "getsockopt(%s)", "IP_FW_ADD");
if (!do_quiet)
show_ipfw(rule);
}
static void
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
zero(int ac, 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
{
int rulenum;
int failed = EX_OK;
av++; ac--;
if (!ac) {
/* clear all entries */
if (setsockopt(s, IPPROTO_IP, IP_FW_ZERO, NULL, 0) < 0)
err(EX_UNAVAILABLE, "setsockopt(%s)", "IP_FW_ZERO");
if (!do_quiet)
printf("Accounting cleared.\n");
return;
}
while (ac) {
/* Rule number */
if (isdigit(**av)) {
rulenum = atoi(*av);
av++;
ac--;
if (setsockopt(s, IPPROTO_IP,
IP_FW_ZERO, &rulenum, sizeof rulenum)) {
warn("rule %u: setsockopt(IP_FW_ZERO)",
rulenum);
failed = EX_UNAVAILABLE;
} else if (!do_quiet)
printf("Entry %d cleared\n", rulenum);
} else {
errx(EX_USAGE, "invalid rule number ``%s''", *av);
}
}
if (failed != EX_OK)
exit(failed);
}
static void
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
resetlog(int ac, 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
{
int rulenum;
int failed = EX_OK;
av++; ac--;
if (!ac) {
/* clear all entries */
if (setsockopt(s, IPPROTO_IP, IP_FW_RESETLOG, NULL, 0) < 0)
err(EX_UNAVAILABLE, "setsockopt(IP_FW_RESETLOG)");
if (!do_quiet)
printf("Logging counts reset.\n");
return;
}
while (ac) {
/* Rule number */
if (isdigit(**av)) {
rulenum = atoi(*av);
av++;
ac--;
if (setsockopt(s, IPPROTO_IP,
IP_FW_RESETLOG, &rulenum, sizeof rulenum)) {
warn("rule %u: setsockopt(IP_FW_RESETLOG)",
rulenum);
failed = EX_UNAVAILABLE;
} else if (!do_quiet)
printf("Entry %d logging count reset\n",
rulenum);
} else {
errx(EX_DATAERR, "invalid rule number ``%s''", *av);
}
}
if (failed != EX_OK)
exit(failed);
}
static void
flush()
{
int cmd = do_pipe ? IP_DUMMYNET_FLUSH : IP_FW_FLUSH;
if (!do_force && !do_quiet) { /* need to ask user */
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;
}
if (setsockopt(s, IPPROTO_IP, cmd, NULL, 0) < 0)
err(EX_UNAVAILABLE, "setsockopt(IP_%s_FLUSH)",
do_pipe ? "DUMMYNET" : "FW");
if (!do_quiet)
printf("Flushed all %s.\n", do_pipe ? "pipes" : "rules");
}
static int
ipfw_main(int ac, char **av)
{
int ch;
if (ac == 1)
show_usage();
/* Set the force flag for non-interactive processes */
do_force = !isatty(STDIN_FILENO);
optind = optreset = 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
while ((ch = getopt(ac, av, "hs:adefNqStv")) != -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
switch (ch) {
case 'h': /* help */
help();
break; /* NOTREACHED */
case 's': /* sort */
do_sort = atoi(optarg);
break;
case 'a':
do_acct = 1;
break;
case 'd':
do_dynamic = 1;
break;
case 'e':
do_expired = 1;
break;
case 'f':
do_force = 1;
break;
case 'N':
do_resolv = 1;
break;
case 'q':
do_quiet = 1;
break;
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
case 'S':
show_sets = 1;
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 't':
do_time = 1;
break;
case 'v': /* verbose */
verbose++;
break;
default:
show_usage();
}
ac -= optind;
av += optind;
NEED1("bad arguments, for usage summary ``ipfw''");
/*
* optional: pipe or queue
*/
if (!strncmp(*av, "pipe", strlen(*av))) {
do_pipe = 1;
ac--;
av++;
} else if (!strncmp(*av, "queue", strlen(*av))) {
do_pipe = 2;
ac--;
av++;
}
NEED1("missing command");
/*
* for pipes and queues we normally say 'pipe NN config'
* but the code is easier to parse as 'pipe config NN'
* so we swap the two arguments.
*/
if (do_pipe > 0 && ac > 1 && *av[0] >= '0' && *av[0] <= '9') {
char *p = av[0];
av[0] = av[1];
av[1] = p;
}
if (!strncmp(*av, "add", strlen(*av)))
add(ac, av);
else if (do_pipe && !strncmp(*av, "config", strlen(*av)))
config_pipe(ac, 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 if (!strncmp(*av, "delete", strlen(*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
delete(ac, av);
else if (!strncmp(*av, "flush", strlen(*av)))
flush();
else if (!strncmp(*av, "zero", strlen(*av)))
zero(ac, av);
else if (!strncmp(*av, "resetlog", strlen(*av)))
resetlog(ac, av);
else if (!strncmp(*av, "print", strlen(*av)) ||
!strncmp(*av, "list", strlen(*av)))
list(ac, 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 if (!strncmp(*av, "set", strlen(*av)))
sets_handler(ac, 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 if (!strncmp(*av, "show", strlen(*av))) {
do_acct++;
list(ac, av);
} else
errx(EX_USAGE, "bad command `%s'", *av);
return 0;
}
static void
ipfw_readfile(int ac, char *av[])
{
#define MAX_ARGS 32
#define WHITESP " \t\f\v\n\r"
char buf[BUFSIZ];
char *a, *p, *args[MAX_ARGS], *cmd = NULL;
char linename[10];
int i=0, lineno=0, qflag=0, pflag=0, status;
FILE *f = NULL;
pid_t preproc = 0;
int c;
while ((c = getopt(ac, av, "D:U:p:q")) != -1)
switch(c) {
case 'D':
if (!pflag)
errx(EX_USAGE, "-D requires -p");
if (i > MAX_ARGS - 2)
errx(EX_USAGE,
"too many -D or -U options");
args[i++] = "-D";
args[i++] = optarg;
break;
case 'U':
if (!pflag)
errx(EX_USAGE, "-U requires -p");
if (i > MAX_ARGS - 2)
errx(EX_USAGE,
"too many -D or -U options");
args[i++] = "-U";
args[i++] = optarg;
break;
case 'p':
pflag = 1;
cmd = optarg;
args[0] = cmd;
i = 1;
break;
case 'q':
qflag = 1;
break;
default:
errx(EX_USAGE, "bad arguments, for usage"
" summary ``ipfw''");
}
av += optind;
ac -= optind;
if (ac != 1)
errx(EX_USAGE, "extraneous filename arguments");
if ((f = fopen(av[0], "r")) == NULL)
err(EX_UNAVAILABLE, "fopen: %s", av[0]);
if (pflag) {
/* pipe through preprocessor (cpp or m4) */
int pipedes[2];
args[i] = 0;
if (pipe(pipedes) == -1)
err(EX_OSERR, "cannot create pipe");
switch((preproc = fork())) {
case -1:
err(EX_OSERR, "cannot fork");
case 0:
/* child */
if (dup2(fileno(f), 0) == -1
|| dup2(pipedes[1], 1) == -1)
err(EX_OSERR, "dup2()");
fclose(f);
close(pipedes[1]);
close(pipedes[0]);
execvp(cmd, args);
err(EX_OSERR, "execvp(%s) failed", cmd);
default:
/* parent */
fclose(f);
close(pipedes[1]);
if ((f = fdopen(pipedes[0], "r")) == NULL) {
int savederrno = errno;
(void)kill(preproc, SIGTERM);
errno = savederrno;
err(EX_OSERR, "fdopen()");
}
}
}
while (fgets(buf, BUFSIZ, f)) {
lineno++;
sprintf(linename, "Line %d", lineno);
args[0] = linename;
if (*buf == '#')
continue;
if ((p = strchr(buf, '#')) != NULL)
*p = '\0';
i = 1;
if (qflag)
args[i++] = "-q";
for (a = strtok(buf, WHITESP);
a && i < MAX_ARGS; a = strtok(NULL, WHITESP), i++)
args[i] = a;
if (i == (qflag? 2: 1))
continue;
if (i == MAX_ARGS)
errx(EX_USAGE, "%s: too many arguments",
linename);
args[i] = NULL;
ipfw_main(i, args);
}
fclose(f);
if (pflag) {
if (waitpid(preproc, &status, 0) == -1)
errx(EX_OSERR, "waitpid()");
if (WIFEXITED(status) && WEXITSTATUS(status) != EX_OK)
errx(EX_UNAVAILABLE,
"preprocessor exited with status %d",
WEXITSTATUS(status));
else if (WIFSIGNALED(status))
errx(EX_UNAVAILABLE,
"preprocessor exited with signal %d",
WTERMSIG(status));
}
}
int
main(int ac, char *av[])
{
s = socket(AF_INET, SOCK_RAW, IPPROTO_RAW);
if (s < 0)
err(EX_UNAVAILABLE, "socket");
/*
* If the last argument is an absolute pathname, interpret it
* as a file to be preprocessed.
*/
if (ac > 1 && av[ac - 1][0] == '/' && access(av[ac - 1], R_OK) == 0)
ipfw_readfile(ac, av);
else
ipfw_main(ac, av);
return EX_OK;
}
Reference in New Issue Copy Permalink