d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
95641278ef
freebsd-dev/usr.sbin/rarpd/rarpd.c

1055 lines
24 KiB
C
Raw Normal View History

Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
/*
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
* Copyright (c) 1990, 1991, 1992, 1993, 1996
* The Regents of the University of California. All rights reserved.
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that: (1) source code distributions
* retain the above copyright notice and this paragraph in its entirety, (2)
* distributions including binary code include the above copyright notice and
* this paragraph in its entirety in the documentation or other materials
* provided with the distribution, and (3) all advertising materials mentioning
* features or use of this software display the following acknowledgement:
* ``This product includes software developed by the University of California,
* Lawrence Berkeley Laboratory and its contributors.'' Neither the name of
* the University nor the names of its contributors may be used to endorse
* or promote products derived from this software without specific prior
* written permission.
* THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
*/
Mdoc'ify man page.
1997-10-13 11:03:36 +00:00
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#ifndef lint
Mdoc'ify man page.
1997-10-13 11:03:36 +00:00
static const char copyright[] =
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
"@(#) Copyright (c) 1990, 1991, 1992, 1993, 1996\n\
The Regents of the University of California. All rights reserved.\n";
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#endif /* not lint */
#ifndef lint
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
static const char rcsid[] =
$Id$ -> $FreeBSD$
1999-08-28 01:35:59 +00:00
"$FreeBSD$";
Mdoc'ify man page.
1997-10-13 11:03:36 +00:00
#endif /* not lint */
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
/*
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
* rarpd - Reverse ARP Daemon
*
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
* Usage: rarpd -a [ -dfsv ] [ hostname ]
* rarpd [ -dfsv ] interface [ hostname ]
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
*
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
* 'hostname' is optional solely for backwards compatibility with Sun's rarpd.
* Currently, the argument is ignored.
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#include <sys/param.h>
#include <sys/file.h>
#include <sys/ioctl.h>
#include <sys/socket.h>
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#include <sys/time.h>
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#include <net/bpf.h>
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
#include <net/ethernet.h>
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#include <net/if.h>
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#include <net/if_types.h>
#include <net/if_dl.h>
#include <net/route.h>
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#include <netinet/in.h>
#include <netinet/if_ether.h>
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#include <arpa/inet.h>
#include <errno.h>
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#include <netdb.h>
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
#include <stdarg.h>
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#include <stdio.h>
#include <string.h>
#include <syslog.h>
#include <stdlib.h>
#include <unistd.h>
Eliminated includes of the "temporary" backwards compatibility header <sys/dir.h> in applications. Maintained existing (inadequate) ifdefs for dir.h vs dirent.h in libdialog, amd and rarpd, but didn't add any new ones.
1996-09-24 08:08:11 +00:00
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
#if defined(SUNOS4) || defined(__FreeBSD__) /* XXX */
#define HAVE_DIRENT_H
#endif
#ifdef HAVE_DIRENT_H
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#include <dirent.h>
#else
#include <sys/dir.h>
#endif
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
/* Cast a struct sockaddr to a structaddr_in */
#define SATOSIN(sa) ((struct sockaddr_in *)(sa))
#ifndef TFTP_DIR
#define TFTP_DIR "/tftpboot"
#endif
#if BSD >= 199200
#define ARPSECS (20 * 60) /* as per code in netinet/if_ether.c */
#define REVARP_REQUEST ARPOP_REVREQUEST
#define REVARP_REPLY ARPOP_REVREPLY
#endif
#ifndef ETHERTYPE_REVARP
#define ETHERTYPE_REVARP 0x8035
#define REVARP_REQUEST 3
#define REVARP_REPLY 4
#endif
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
/*
* Map field names in ether_arp struct. What a pain in the neck.
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#ifdef SUNOS3
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#undef arp_sha
#undef arp_spa
#undef arp_tha
#undef arp_tpa
#define arp_sha arp_xsha
#define arp_spa arp_xspa
#define arp_tha arp_xtha
#define arp_tpa arp_xtpa
#endif
/*
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
* The structure for each interface.
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
*/
struct if_info {
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
struct if_info *ii_next;
int ii_fd; /* BPF file descriptor */
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
u_long ii_ipaddr; /* IP address of this interface */
u_long ii_netmask; /* subnet or net mask */
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
u_char ii_eaddr[6]; /* Ethernet address of this interface */
char ii_ifname[sizeof(((struct ifreq *)0)->ifr_name) + 1];
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
};
/*
* The list of all interfaces that are being listened to. rarp_loop()
* "selects" on the descriptors in this list.
*/
struct if_info *iflist;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
int verbose; /* verbose messages */
int s; /* inet datagram socket */
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
const char *tftp_dir = TFTP_DIR; /* tftp directory */
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
int dflag; /* messages to stdout/stderr, not syslog(3) */
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
int sflag; /* ignore /tftpboot */
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
static u_char zero[6];
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
static int bpf_open(void);
static u_long choose_ipaddr(u_long **, u_long, u_long);
static char *eatoa(u_char *);
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
static int expand_syslog_m(const char *fmt, char **newfmt);
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
static void init(char *);
static void init_one(struct ifreq *, char *);
static char *intoa(u_long);
static u_long ipaddrtonetmask(u_long);
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
static void logmsg(int, const char *, ...) __printflike(2, 3);
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
static int rarp_bootable(u_long);
static int rarp_check(u_char *, u_int);
static void rarp_loop(void);
static int rarp_open(char *);
static void rarp_process(struct if_info *, u_char *, u_int);
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
static void rarp_reply(struct if_info *, struct ether_header *,
u_long, u_int);
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
static void update_arptab(u_char *, u_long);
static void usage(void);
Add -s flag to always supply mapping if known, ignoring the presence or absence of files in /tftpboot.
1996-11-27 20:45:10 +00:00
Fix warning: return type of `main' is not `int'
1999-09-15 01:58:44 +00:00
int
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
main(int argc, char *argv[])
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
int op;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
char *ifname, *hostname, *name;
int aflag = 0; /* listen on "all" interfaces */
int fflag = 0; /* don't fork */
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if ((name = strrchr(argv[0], '/')) != NULL)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
++name;
else
name = argv[0];
if (*name == '-')
++name;
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
/*
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
* All error reporting is done through syslog, unless -d is specified
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
openlog(name, LOG_PID | LOG_CONS, LOG_DAEMON);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
opterr = 0;
o Allow rarpd to accept an additional '-t directory' argument, specifying an alternative to /tftpboot. This is useful it you're using tftpd with an alternative root (using -s), and would like rarpd to respond selectively to RARP requests using the same criteria as tftp. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2001-11-16 16:49:21 +00:00
while ((op = getopt(argc, argv, "adfst:v")) != -1) {
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
switch (op) {
case 'a':
++aflag;
break;
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
case 'd':
++dflag;
break;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
case 'f':
++fflag;
break;
Add -s flag to always supply mapping if known, ignoring the presence or absence of files in /tftpboot.
1996-11-27 20:45:10 +00:00
case 's':
++sflag;
break;
o Allow rarpd to accept an additional '-t directory' argument, specifying an alternative to /tftpboot. This is useful it you're using tftpd with an alternative root (using -s), and would like rarpd to respond selectively to RARP requests using the same criteria as tftp. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2001-11-16 16:49:21 +00:00
case 't':
tftp_dir = optarg;
break;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
case 'v':
++verbose;
break;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
default:
usage();
/* NOTREACHED */
}
}
ifname = argv[optind++];
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
hostname = ifname ? argv[optind] : NULL;
if ((aflag && ifname) || (!aflag && ifname == NULL))
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
usage();
if (aflag)
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
init(NULL);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
else
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
init(ifname);
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (!fflag) {
Use daemon() to deamonify ourselves.
1995-07-18 21:35:32 +00:00
if (daemon(0,0)) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "cannot fork");
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
exit(1);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
}
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
rarp_loop();
Remove incorrect section name. Incomplete -Wall cleaning.
1999-11-27 17:06:40 +00:00
return(0);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
/*
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
* Add to the interface list.
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
*/
void
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
init_one(struct ifreq *ifrp, char *target)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
struct if_info *ii;
struct sockaddr_dl *ll;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
int family;
struct ifreq ifr;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
family = ifrp->ifr_addr.sa_family;
switch (family) {
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
case AF_INET:
#if BSD >= 199100
case AF_LINK:
#endif
(void)strncpy(ifr.ifr_name, ifrp->ifr_name,
sizeof(ifrp->ifr_name));
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(s, SIOCGIFFLAGS, (char *)&ifr) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR,
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
"SIOCGIFFLAGS: %.*s: %m",
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
(int)sizeof(ifrp->ifr_name), ifrp->ifr_name);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
exit(1);
}
if ((ifr.ifr_flags & IFF_UP) == 0 ||
(ifr.ifr_flags & (IFF_LOOPBACK | IFF_POINTOPOINT)) != 0)
return;
break;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
default:
return;
}
/* Don't bother going any further if not the target interface */
if (target != NULL &&
strncmp(ifrp->ifr_name, target, sizeof(ifrp->ifr_name)) != 0)
return;
/* Look for interface in list */
for (ii = iflist; ii != NULL; ii = ii->ii_next)
if (strncmp(ifrp->ifr_name, ii->ii_ifname,
sizeof(ifrp->ifr_name)) == 0)
break;
/* Allocate a new one if not found */
if (ii == NULL) {
ii = (struct if_info *)malloc(sizeof(*ii));
if (ii == NULL) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "malloc: %m");
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
exit(1);
}
bzero(ii, sizeof(*ii));
ii->ii_fd = -1;
(void)strncpy(ii->ii_ifname, ifrp->ifr_name,
sizeof(ifrp->ifr_name));
ii->ii_ifname[sizeof(ii->ii_ifname) - 1] = '\0';
ii->ii_next = iflist;
iflist = ii;
}
switch (family) {
case AF_INET:
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(s, SIOCGIFADDR, (char *)&ifr) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "ipaddr SIOCGIFADDR: %s: %m",
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
ii->ii_ifname);
exit(1);
}
ii->ii_ipaddr = SATOSIN(&ifr.ifr_addr)->sin_addr.s_addr;
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(s, SIOCGIFNETMASK, (char *)&ifr) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "SIOCGIFNETMASK: %m");
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
exit(1);
}
ii->ii_netmask = SATOSIN(&ifr.ifr_addr)->sin_addr.s_addr;
if (ii->ii_netmask == 0)
ii->ii_netmask = ipaddrtonetmask(ii->ii_ipaddr);
if (ii->ii_fd < 0) {
ii->ii_fd = rarp_open(ii->ii_ifname);
#if BSD < 199100
/* Use BPF descriptor to get ethernet address. */
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(ii->ii_fd, SIOCGIFADDR, (char *)&ifr) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "eaddr SIOCGIFADDR: %s: %m",
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
ii->ii_ifname);
exit(1);
}
bcopy(&ifr.ifr_addr.sa_data[0], ii->ii_eaddr, 6);
#endif
}
break;
#if BSD >= 199100
case AF_LINK:
ll = (struct sockaddr_dl *)&ifrp->ifr_addr;
if (ll->sdl_type == IFT_ETHER)
bcopy(LLADDR(ll), ii->ii_eaddr, 6);
break;
#endif
}
}
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
/*
* Initialize all "candidate" interfaces that are in the system
* configuration list. A "candidate" is up, not loopback and not
* point to point.
*/
void
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
init(char *target)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
Fixed world breakage on systems where ntohl() doesn't return u_long (e.g., on alphas, or even on i386's with a POSIX-200x-conformant ntohl() (ntohl() returns uint32_t which is u_int on i386's)). Fixed related bugs and bogons while I'm here: - ntohl() was "fixed" for printing in 1 place by casting to "(unsigned int )". This breaks the value on systems where u_int is smaller than uint32_t, and has 2 style bugs. - spell u_int consistently (never use "unsigned"). - break K&R support some more (don't cast malloc()'s arg to a wrong type...).
2001-06-13 11:56:00 +00:00
u_int n;
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
struct ifreq *ifrp, *ifend;
struct if_info *ii, *nii, *lii;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
struct ifconf ifc;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
struct ifreq ibuf[16];
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if ((s = socket(AF_INET, SOCK_DGRAM, 0)) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "socket: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
ifc.ifc_len = sizeof ibuf;
ifc.ifc_buf = (caddr_t)ibuf;
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if ((ioctl(s, SIOCGIFCONF, (char *)&ifc) == -1) ||
((u_int)ifc.ifc_len < sizeof(struct ifreq))) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "SIOCGIFCONF: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
ifrp = ibuf;
ifend = (struct ifreq *)((char *)ibuf + ifc.ifc_len);
while (ifrp < ifend) {
init_one(ifrp, target);
#if BSD >= 199100
n = ifrp->ifr_addr.sa_len + sizeof(ifrp->ifr_name);
if (n < sizeof(*ifrp))
n = sizeof(*ifrp);
ifrp = (struct ifreq *)((char *)ifrp + n);
#else
++ifrp;
#endif
}
/* Throw away incomplete interfaces */
lii = NULL;
for (ii = iflist; ii != NULL; ii = nii) {
nii = ii->ii_next;
if (ii->ii_ipaddr == 0 ||
bcmp(ii->ii_eaddr, zero, 6) == 0) {
if (lii == NULL)
iflist = nii;
else
lii->ii_next = nii;
if (ii->ii_fd >= 0)
close(ii->ii_fd);
free(ii);
continue;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
lii = ii;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
/* Verbose stuff */
if (verbose)
for (ii = iflist; ii != NULL; ii = ii->ii_next)
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_DEBUG, "%s %s 0x%08lx %s",
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
ii->ii_ifname, intoa(ntohl(ii->ii_ipaddr)),
Fixed world breakage on systems where ntohl() doesn't return u_long (e.g., on alphas, or even on i386's with a POSIX-200x-conformant ntohl() (ntohl() returns uint32_t which is u_int on i386's)). Fixed related bugs and bogons while I'm here: - ntohl() was "fixed" for printing in 1 place by casting to "(unsigned int )". This breaks the value on systems where u_int is smaller than uint32_t, and has 2 style bugs. - spell u_int consistently (never use "unsigned"). - break K&R support some more (don't cast malloc()'s arg to a wrong type...).
2001-06-13 11:56:00 +00:00
(u_long)ntohl(ii->ii_netmask), eatoa(ii->ii_eaddr));
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
void
usage(void)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
(void)fprintf(stderr, "usage: rarpd [-adfsv] [interface]\n");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
int
bpf_open(void)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
int fd;
int n = 0;
char device[sizeof "/dev/bpf000"];
/*
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
* Go through all the minors and find one that isn't in use.
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
*/
do {
(void)sprintf(device, "/dev/bpf%d", n++);
fd = open(device, O_RDWR);
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
} while ((fd == -1) && (errno == EBUSY));
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (fd == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "%s: %m", device);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
exit(1);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
return fd;
}
/*
* Open a BPF file and attach it to the interface named 'device'.
* Set immediate mode, and set a filter that accepts only RARP requests.
*/
int
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
rarp_open(char *device)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
int fd;
struct ifreq ifr;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
u_int dlt;
int immediate;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
static struct bpf_insn insns[] = {
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
BPF_STMT(BPF_LD|BPF_H|BPF_ABS, 12),
BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, ETHERTYPE_REVARP, 0, 3),
BPF_STMT(BPF_LD|BPF_H|BPF_ABS, 20),
BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, REVARP_REQUEST, 0, 1),
BPF_STMT(BPF_RET|BPF_K, sizeof(struct ether_arp) +
sizeof(struct ether_header)),
BPF_STMT(BPF_RET|BPF_K, 0),
};
static struct bpf_program filter = {
sizeof insns / sizeof(insns[0]),
insns
};
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
fd = bpf_open();
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
/*
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
* Set immediate mode so packets are processed as they arrive.
*/
immediate = 1;
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(fd, BIOCIMMEDIATE, &immediate) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "BIOCIMMEDIATE: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
(void)strncpy(ifr.ifr_name, device, sizeof ifr.ifr_name);
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(fd, BIOCSETIF, (caddr_t)&ifr) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "BIOCSETIF: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
/*
* Check that the data link layer is an Ethernet; this code won't
* work with anything else.
*/
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(fd, BIOCGDLT, (caddr_t)&dlt) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "BIOCGDLT: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (dlt != DLT_EN10MB) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "%s is not an ethernet", device);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
/*
* Set filter program.
*/
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(fd, BIOCSETF, (caddr_t)&filter) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "BIOCSETF: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
return fd;
}
/*
* Perform various sanity checks on the RARP request packet. Return
* false on failure and log the reason.
*/
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
int
rarp_check(u_char *p, u_int len)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
struct ether_header *ep = (struct ether_header *)p;
struct ether_arp *ap = (struct ether_arp *)(p + sizeof(*ep));
if (len < sizeof(*ep) + sizeof(*ap)) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "truncated request, got %u, expected %lu",
Comply with POSIX rules: o Use %u for printing u_int. o Cast sizeof() to u_long, and print with %lu Reviewed by: wollman
2001-06-16 18:55:13 +00:00
len, (u_long)(sizeof(*ep) + sizeof(*ap)));
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
return 0;
}
/*
* XXX This test might be better off broken out...
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (ntohs(ep->ether_type) != ETHERTYPE_REVARP ||
ntohs(ap->arp_hrd) != ARPHRD_ETHER ||
ntohs(ap->arp_op) != REVARP_REQUEST ||
ntohs(ap->arp_pro) != ETHERTYPE_IP ||
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
ap->arp_hln != 6 || ap->arp_pln != 4) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_DEBUG, "request fails sanity check");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
return 0;
}
if (bcmp((char *)&ep->ether_shost, (char *)&ap->arp_sha, 6) != 0) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_DEBUG, "ether/arp sender address mismatch");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
return 0;
}
if (bcmp((char *)&ap->arp_sha, (char *)&ap->arp_tha, 6) != 0) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_DEBUG, "ether/arp target address mismatch");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
return 0;
}
return 1;
}
#ifndef FD_SETSIZE
#define FD_SET(n, fdp) ((fdp)->fds_bits[0] |= (1 << (n)))
#define FD_ISSET(n, fdp) ((fdp)->fds_bits[0] & (1 << (n)))
#define FD_ZERO(fdp) ((fdp)->fds_bits[0] = 0)
#endif
/*
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
* Loop indefinitely listening for RARP requests on the
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
* interfaces in 'iflist'.
*/
void
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
rarp_loop(void)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
u_char *buf, *bp, *ep;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
int cc, fd;
fd_set fds, listeners;
int bufsize, maxfd = 0;
struct if_info *ii;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (iflist == NULL) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "no interfaces");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(iflist->ii_fd, BIOCGBLEN, (caddr_t)&bufsize) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "BIOCGBLEN: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
Fixed world breakage on systems where ntohl() doesn't return u_long (e.g., on alphas, or even on i386's with a POSIX-200x-conformant ntohl() (ntohl() returns uint32_t which is u_int on i386's)). Fixed related bugs and bogons while I'm here: - ntohl() was "fixed" for printing in 1 place by casting to "(unsigned int )". This breaks the value on systems where u_int is smaller than uint32_t, and has 2 style bugs. - spell u_int consistently (never use "unsigned"). - break K&R support some more (don't cast malloc()'s arg to a wrong type...).
2001-06-13 11:56:00 +00:00
buf = malloc(bufsize);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (buf == NULL) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "malloc: %m");
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
exit(1);
}
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
while (1) {
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
/*
* Find the highest numbered file descriptor for select().
* Initialize the set of descriptors to listen to.
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
FD_ZERO(&fds);
for (ii = iflist; ii != NULL; ii = ii->ii_next) {
FD_SET(ii->ii_fd, &fds);
if (ii->ii_fd > maxfd)
maxfd = ii->ii_fd;
}
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
listeners = fds;
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (select(maxfd + 1, &listeners, NULL, NULL, NULL) == -1) {
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
/* Don't choke when we get ptraced */
if (errno == EINTR)
continue;
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "select: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
for (ii = iflist; ii != NULL; ii = ii->ii_next) {
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
fd = ii->ii_fd;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (!FD_ISSET(fd, &listeners))
continue;
again:
cc = read(fd, (char *)buf, bufsize);
/* Don't choke when we get ptraced */
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if ((cc == -1) && (errno == EINTR))
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
goto again;
#if defined(SUNOS3) || defined(SUNOS4)
/*
* Due to a SunOS bug, after 2^31 bytes, the
* file offset overflows and read fails with
* EINVAL. The lseek() to 0 will fix things.
*/
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (cc == -1) {
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (errno == EINVAL &&
(long)(tell(fd) + bufsize) < 0) {
(void)lseek(fd, 0, 0);
goto again;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "read: %m");
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
exit(1);
}
#endif
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
/* Loop through the packet(s) */
#define bhp ((struct bpf_hdr *)bp)
bp = buf;
ep = bp + cc;
while (bp < ep) {
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
u_int caplen, hdrlen;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
caplen = bhp->bh_caplen;
hdrlen = bhp->bh_hdrlen;
if (rarp_check(bp + hdrlen, caplen))
Make the man page reflect reality. Add BUGS section about DNS. Logging cleanups (including logging the requestor's MAC address instead of the server's).
1996-11-27 20:42:09 +00:00
rarp_process(ii, bp + hdrlen, caplen);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
bp += BPF_WORDALIGN(hdrlen + caplen);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
}
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#undef bhp
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
/*
* True if this server can boot the host whose IP address is 'addr'.
* This check is made by looking in the tftp directory for the
* configuration file.
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
int
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
rarp_bootable(u_long addr)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
#ifdef HAVE_DIRENT_H
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
struct dirent *dent;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#else
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
struct direct *dent;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
#endif
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
DIR *d;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
char ipname[9];
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
static DIR *dd = NULL;
Fixed world breakage on systems where ntohl() doesn't return u_long (e.g., on alphas, or even on i386's with a POSIX-200x-conformant ntohl() (ntohl() returns uint32_t which is u_int on i386's)). Fixed related bugs and bogons while I'm here: - ntohl() was "fixed" for printing in 1 place by casting to "(unsigned int )". This breaks the value on systems where u_int is smaller than uint32_t, and has 2 style bugs. - spell u_int consistently (never use "unsigned"). - break K&R support some more (don't cast malloc()'s arg to a wrong type...).
2001-06-13 11:56:00 +00:00
(void)sprintf(ipname, "%08lX", (u_long)ntohl(addr));
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
/*
* If directory is already open, rewind it. Otherwise, open it.
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if ((d = dd) != NULL)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
rewinddir(d);
else {
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (chdir(tftp_dir) == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "chdir: %s: %m", tftp_dir);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
d = opendir(".");
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (d == NULL) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "opendir: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
exit(1);
}
dd = d;
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
while ((dent = readdir(d)) != NULL)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
if (strncmp(dent->d_name, ipname, 8) == 0)
return 1;
return 0;
}
/*
* Given a list of IP addresses, 'alist', return the first address that
* is on network 'net'; 'netmask' is a mask indicating the network portion
* of the address.
*/
u_long
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
choose_ipaddr(u_long **alist, u_long net, u_long netmask)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
for (; *alist; ++alist)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
if ((**alist & netmask) == net)
return **alist;
return 0;
}
/*
* Answer the RARP request in 'pkt', on the interface 'ii'. 'pkt' has
* already been checked for validity. The reply is overlaid on the request.
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
void
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
rarp_process(struct if_info *ii, u_char *pkt, u_int len)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
struct ether_header *ep;
struct hostent *hp;
u_long target_ipaddr;
char ename[256];
ep = (struct ether_header *)pkt;
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
/* should this be arp_tha? */
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ether_ntohost(ename, (struct ether_addr *)&ep->ether_shost) != 0) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "cannot map %s to name",
Make the man page reflect reality. Add BUGS section about DNS. Logging cleanups (including logging the requestor's MAC address instead of the server's).
1996-11-27 20:42:09 +00:00
eatoa(ep->ether_shost));
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
return;
Make the man page reflect reality. Add BUGS section about DNS. Logging cleanups (including logging the requestor's MAC address instead of the server's).
1996-11-27 20:42:09 +00:00
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
Make the man page reflect reality. Add BUGS section about DNS. Logging cleanups (including logging the requestor's MAC address instead of the server's).
1996-11-27 20:42:09 +00:00
if ((hp = gethostbyname(ename)) == NULL) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "cannot map %s to IP address", ename);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
return;
Make the man page reflect reality. Add BUGS section about DNS. Logging cleanups (including logging the requestor's MAC address instead of the server's).
1996-11-27 20:42:09 +00:00
}
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
/*
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
* Choose correct address from list.
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (hp->h_addrtype != AF_INET) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "cannot handle non IP addresses for %s",
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
ename);
return;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
}
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
target_ipaddr = choose_ipaddr((u_long **)hp->h_addr_list,
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
ii->ii_ipaddr & ii->ii_netmask,
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
ii->ii_netmask);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (target_ipaddr == 0) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "cannot find %s on net %s",
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
ename, intoa(ntohl(ii->ii_ipaddr & ii->ii_netmask)));
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
return;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Add -s flag to always supply mapping if known, ignoring the presence or absence of files in /tftpboot.
1996-11-27 20:45:10 +00:00
if (sflag || rarp_bootable(target_ipaddr))
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
rarp_reply(ii, ep, target_ipaddr, len);
else if (verbose > 1)
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_INFO, "%s %s at %s DENIED (not bootable)",
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
ii->ii_ifname,
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
eatoa(ep->ether_shost),
intoa(ntohl(target_ipaddr)));
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
/*
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
* Poke the kernel arp tables with the ethernet/ip address combinataion
* given. When processing a reply, we must do this so that the booting
* host (i.e. the guy running rarpd), won't try to ARP for the hardware
* address of the guy being booted (he cannot answer the ARP).
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#if BSD >= 199200
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
struct sockaddr_inarp sin_inarp = {
sizeof(struct sockaddr_inarp), AF_INET, 0,
{0},
{0},
0, 0
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
};
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
struct sockaddr_dl sin_dl = {
sizeof(struct sockaddr_dl), AF_LINK, 0, IFT_ETHER, 0, 6,
0, "", 0, {0}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
};
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
struct {
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
struct rt_msghdr rthdr;
char rtspace[512];
} rtmsg;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
void
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
update_arptab(u_char *ep, u_long ipaddr)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
int cc;
struct sockaddr_inarp *ar, *ar2;
struct sockaddr_dl *ll, *ll2;
struct rt_msghdr *rt;
int xtype, xindex;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
static pid_t pid;
Fix bug in rarpd: Explanation of the bug: when processing its first request, rarpd opens a routing socket to send requests to the arp table. It keeps that socket open afterwards, while waiting for new RARP requests. Meanwhile, the data received on the routing socket fill up until they are about 8Kbytes in size. Any additional data is lost. When rarpd receives its next RARP request, it tries to access the ARP table via a routing socket call, then waits for the answer to its own request. This answer is lost because the received data is already filled: when looking for the reply, rarpd receives only 8kbytes worth of data, then loops waiting forever. Someone please test it on -STABLE and commit it. We can close the PR when testing on STABLE is done. PR: bin/5669 Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>
1998-04-02 13:20:15 +00:00
int r;
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
static int seq;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
Fix bug in rarpd: Explanation of the bug: when processing its first request, rarpd opens a routing socket to send requests to the arp table. It keeps that socket open afterwards, while waiting for new RARP requests. Meanwhile, the data received on the routing socket fill up until they are about 8Kbytes in size. Any additional data is lost. When rarpd receives its next RARP request, it tries to access the ARP table via a routing socket call, then waits for the answer to its own request. This answer is lost because the received data is already filled: when looking for the reply, rarpd receives only 8kbytes worth of data, then loops waiting forever. Someone please test it on -STABLE and commit it. We can close the PR when testing on STABLE is done. PR: bin/5669 Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>
1998-04-02 13:20:15 +00:00
r = socket(PF_ROUTE, SOCK_RAW, 0);
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (r == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "raw route socket: %m");
Fix bug in rarpd: Explanation of the bug: when processing its first request, rarpd opens a routing socket to send requests to the arp table. It keeps that socket open afterwards, while waiting for new RARP requests. Meanwhile, the data received on the routing socket fill up until they are about 8Kbytes in size. Any additional data is lost. When rarpd receives its next RARP request, it tries to access the ARP table via a routing socket call, then waits for the answer to its own request. This answer is lost because the received data is already filled: when looking for the reply, rarpd receives only 8kbytes worth of data, then loops waiting forever. Someone please test it on -STABLE and commit it. We can close the PR when testing on STABLE is done. PR: bin/5669 Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>
1998-04-02 13:20:15 +00:00
exit(1);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
}
Fix bug in rarpd: Explanation of the bug: when processing its first request, rarpd opens a routing socket to send requests to the arp table. It keeps that socket open afterwards, while waiting for new RARP requests. Meanwhile, the data received on the routing socket fill up until they are about 8Kbytes in size. Any additional data is lost. When rarpd receives its next RARP request, it tries to access the ARP table via a routing socket call, then waits for the answer to its own request. This answer is lost because the received data is already filled: when looking for the reply, rarpd receives only 8kbytes worth of data, then loops waiting forever. Someone please test it on -STABLE and commit it. We can close the PR when testing on STABLE is done. PR: bin/5669 Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>
1998-04-02 13:20:15 +00:00
pid = getpid();
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
ar = &sin_inarp;
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
ar->sin_addr.s_addr = ipaddr;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
ll = &sin_dl;
bcopy(ep, LLADDR(ll), 6);
/* Get the type and interface index */
rt = &rtmsg.rthdr;
bzero(rt, sizeof(rtmsg));
rt->rtm_version = RTM_VERSION;
rt->rtm_addrs = RTA_DST;
rt->rtm_type = RTM_GET;
rt->rtm_seq = ++seq;
ar2 = (struct sockaddr_inarp *)rtmsg.rtspace;
bcopy(ar, ar2, sizeof(*ar));
rt->rtm_msglen = sizeof(*rt) + sizeof(*ar);
errno = 0;
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if ((write(r, rt, rt->rtm_msglen) == -1) && (errno != ESRCH)) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "rtmsg get write: %m");
Fix bug in rarpd: Explanation of the bug: when processing its first request, rarpd opens a routing socket to send requests to the arp table. It keeps that socket open afterwards, while waiting for new RARP requests. Meanwhile, the data received on the routing socket fill up until they are about 8Kbytes in size. Any additional data is lost. When rarpd receives its next RARP request, it tries to access the ARP table via a routing socket call, then waits for the answer to its own request. This answer is lost because the received data is already filled: when looking for the reply, rarpd receives only 8kbytes worth of data, then loops waiting forever. Someone please test it on -STABLE and commit it. We can close the PR when testing on STABLE is done. PR: bin/5669 Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>
1998-04-02 13:20:15 +00:00
close(r);
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
return;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
do {
cc = read(r, rt, sizeof(rtmsg));
} while (cc > 0 && (rt->rtm_seq != seq || rt->rtm_pid != pid));
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (cc == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "rtmsg get read: %m");
Fix bug in rarpd: Explanation of the bug: when processing its first request, rarpd opens a routing socket to send requests to the arp table. It keeps that socket open afterwards, while waiting for new RARP requests. Meanwhile, the data received on the routing socket fill up until they are about 8Kbytes in size. Any additional data is lost. When rarpd receives its next RARP request, it tries to access the ARP table via a routing socket call, then waits for the answer to its own request. This answer is lost because the received data is already filled: when looking for the reply, rarpd receives only 8kbytes worth of data, then loops waiting forever. Someone please test it on -STABLE and commit it. We can close the PR when testing on STABLE is done. PR: bin/5669 Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>
1998-04-02 13:20:15 +00:00
close(r);
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
return;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
ll2 = (struct sockaddr_dl *)((u_char *)ar2 + ar2->sin_len);
if (ll2->sdl_family != AF_LINK) {
/*
* XXX I think this means the ip address is not on a
* directly connected network (the family is AF_INET in
* this case).
*/
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "bogus link family (%d) wrong net for %08lX?\n",
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
ll2->sdl_family, ipaddr);
Fix bug in rarpd: Explanation of the bug: when processing its first request, rarpd opens a routing socket to send requests to the arp table. It keeps that socket open afterwards, while waiting for new RARP requests. Meanwhile, the data received on the routing socket fill up until they are about 8Kbytes in size. Any additional data is lost. When rarpd receives its next RARP request, it tries to access the ARP table via a routing socket call, then waits for the answer to its own request. This answer is lost because the received data is already filled: when looking for the reply, rarpd receives only 8kbytes worth of data, then loops waiting forever. Someone please test it on -STABLE and commit it. We can close the PR when testing on STABLE is done. PR: bin/5669 Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>
1998-04-02 13:20:15 +00:00
close(r);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
return;
}
xtype = ll2->sdl_type;
xindex = ll2->sdl_index;
/* Set the new arp entry */
bzero(rt, sizeof(rtmsg));
rt->rtm_version = RTM_VERSION;
rt->rtm_addrs = RTA_DST | RTA_GATEWAY;
rt->rtm_inits = RTV_EXPIRE;
rt->rtm_rmx.rmx_expire = time(0) + ARPSECS;
rt->rtm_flags = RTF_HOST | RTF_STATIC;
rt->rtm_type = RTM_ADD;
rt->rtm_seq = ++seq;
bcopy(ar, ar2, sizeof(*ar));
ll2 = (struct sockaddr_dl *)((u_char *)ar2 + sizeof(*ar2));
bcopy(ll, ll2, sizeof(*ll));
ll2->sdl_type = xtype;
ll2->sdl_index = xindex;
rt->rtm_msglen = sizeof(*rt) + sizeof(*ar2) + sizeof(*ll2);
errno = 0;
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if ((write(r, rt, rt->rtm_msglen) == -1) && (errno != EEXIST)) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "rtmsg add write: %m");
Fix bug in rarpd: Explanation of the bug: when processing its first request, rarpd opens a routing socket to send requests to the arp table. It keeps that socket open afterwards, while waiting for new RARP requests. Meanwhile, the data received on the routing socket fill up until they are about 8Kbytes in size. Any additional data is lost. When rarpd receives its next RARP request, it tries to access the ARP table via a routing socket call, then waits for the answer to its own request. This answer is lost because the received data is already filled: when looking for the reply, rarpd receives only 8kbytes worth of data, then loops waiting forever. Someone please test it on -STABLE and commit it. We can close the PR when testing on STABLE is done. PR: bin/5669 Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>
1998-04-02 13:20:15 +00:00
close(r);
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
return;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
do {
cc = read(r, rt, sizeof(rtmsg));
} while (cc > 0 && (rt->rtm_seq != seq || rt->rtm_pid != pid));
Fix bug in rarpd: Explanation of the bug: when processing its first request, rarpd opens a routing socket to send requests to the arp table. It keeps that socket open afterwards, while waiting for new RARP requests. Meanwhile, the data received on the routing socket fill up until they are about 8Kbytes in size. Any additional data is lost. When rarpd receives its next RARP request, it tries to access the ARP table via a routing socket call, then waits for the answer to its own request. This answer is lost because the received data is already filled: when looking for the reply, rarpd receives only 8kbytes worth of data, then loops waiting forever. Someone please test it on -STABLE and commit it. We can close the PR when testing on STABLE is done. PR: bin/5669 Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>
1998-04-02 13:20:15 +00:00
close(r);
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (cc == -1) {
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "rtmsg add read: %m");
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
return;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#else
void
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
update_arptab(u_char *ep, u_long ipaddr)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
struct arpreq request;
struct sockaddr_in *sin;
request.arp_flags = 0;
sin = (struct sockaddr_in *)&request.arp_pa;
sin->sin_family = AF_INET;
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
sin->sin_addr.s_addr = ipaddr;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
request.arp_ha.sa_family = AF_UNSPEC;
bcopy((char *)ep, (char *)request.arp_ha.sa_data, 6);
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
if (ioctl(s, SIOCSARP, (caddr_t)&request) == -1)
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "SIOCSARP: %m");
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#endif
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
/*
* Build a reverse ARP packet and sent it out on the interface.
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
* 'ep' points to a valid REVARP_REQUEST. The REVARP_REPLY is built
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
* on top of the request, then written to the network.
*
* RFC 903 defines the ether_arp fields as follows. The following comments
* are taken (more or less) straight from this document.
*
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
* REVARP_REQUEST
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
*
* arp_sha is the hardware address of the sender of the packet.
* arp_spa is undefined.
* arp_tha is the 'target' hardware address.
* In the case where the sender wishes to determine his own
* protocol address, this, like arp_sha, will be the hardware
* address of the sender.
* arp_tpa is undefined.
*
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
* REVARP_REPLY
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
*
* arp_sha is the hardware address of the responder (the sender of the
* reply packet).
* arp_spa is the protocol address of the responder (see the note below).
* arp_tha is the hardware address of the target, and should be the same as
* that which was given in the request.
* arp_tpa is the protocol address of the target, that is, the desired address.
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
*
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
* Note that the requirement that arp_spa be filled in with the responder's
Remove trailing whitespace.
1995-05-30 03:57:47 +00:00
* protocol is purely for convenience. For instance, if a system were to use
* both ARP and RARP, then the inclusion of the valid protocol-hardware
* address pair (arp_spa, arp_sha) may eliminate the need for a subsequent
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
* ARP request.
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
void
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
rarp_reply(struct if_info *ii, struct ether_header *ep, u_long ipaddr,
u_int len)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
u_int n;
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
struct ether_arp *ap = (struct ether_arp *)(ep + 1);
update_arptab((u_char *)&ap->arp_sha, ipaddr);
/*
* Build the rarp reply by modifying the rarp request in place.
*/
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
ap->arp_op = htons(REVARP_REPLY);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
#ifdef BROKEN_BPF
ep->ether_type = ETHERTYPE_REVARP;
#endif
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
bcopy((char *)&ap->arp_sha, (char *)&ep->ether_dhost, 6);
bcopy((char *)ii->ii_eaddr, (char *)&ep->ether_shost, 6);
bcopy((char *)ii->ii_eaddr, (char *)&ap->arp_sha, 6);
bcopy((char *)&ipaddr, (char *)ap->arp_tpa, 4);
/* Target hardware is unchanged. */
bcopy((char *)&ii->ii_ipaddr, (char *)ap->arp_spa, 4);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
/* Zero possible garbage after packet. */
bzero((char *)ep + (sizeof(*ep) + sizeof(*ap)),
len - (sizeof(*ep) + sizeof(*ap)));
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
n = write(ii->ii_fd, (char *)ep, len);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (n != len)
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_ERR, "write: only %d of %d bytes written", n, len);
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
if (verbose)
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_INFO, "%s %s at %s REPLIED", ii->ii_ifname,
Make the man page reflect reality. Add BUGS section about DNS. Logging cleanups (including logging the requestor's MAC address instead of the server's).
1996-11-27 20:42:09 +00:00
eatoa(ap->arp_tha),
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
intoa(ntohl(ipaddr)));
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
/*
* Get the netmask of an IP address. This routine is used if
* SIOCGIFNETMASK doesn't work.
*/
u_long
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
ipaddrtonetmask(u_long addr)
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
{
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
addr = ntohl(addr);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
if (IN_CLASSA(addr))
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
return htonl(IN_CLASSA_NET);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
if (IN_CLASSB(addr))
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
return htonl(IN_CLASSB_NET);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
if (IN_CLASSC(addr))
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
return htonl(IN_CLASSC_NET);
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
logmsg(LOG_DEBUG, "unknown IP address class: %08lX", addr);
Although I got rarpd to work, it was largely through kludgery. Bill Fenner was kind enough to point out the error of my ways. This incorporates diffs from him which: - Keep everything in network order. - Log the booted ether & ip address, instead of my address on that net - change several exit()'s to return()'s, so that rarpd continues running even if it thinks it's in a weird state. One small tweak by me: in rarp_bootable(), we have to make sure to construct 'ipname' in host byte order (if we don't, we have to specify /tftpboot/<remote IP in hex> with <remote IP in hex> in network byte order, which is confusing). Also restored use of <dirent.h> rather than <sys/dir.h> as pointed out by bde. Also updated the man page so that the -v flag is documented. With any luck, I won't have to touch this thing again.
1996-11-19 23:57:06 +00:00
return htonl(0xffffffff);
Obtained from: An old BPF release packaged with the tcpdump-2.0 source code. "Yes Virginia, there is a rarpd." (Before anyone asks, this *not* the rarpd from NetBSD. It did come from the same place as theirs, however.) This is a port of the rarpd program included with the tcpdump-2.0 source code (which I finally unearthed after scrounging around some of the darker corners of the Internet). It's as close to the original as I could keep it except for the following changes: - The original program was based on an older version of the Berkeley Packet Filter which used different filter programming instructions. Fortunately, an updated RARP packet filter is available right in the BPF man page so this was easy to fix. - The old code didn't know how to deal with variable length addresses in ifreq buffers. This has been fixed. - Some byte order weirdness had to be fixed. The sanity checks in rarp_check() needed some htons()es, and the rarp_reply() function needed to properly set the ether_type field in the ethernet header to ETHERTYPE_REVARP before transmitting the packet, otherwise the bytes in ether_type would wind up reversed. It is important to note that using htons(ETHERTYPE_REVARP) will not work. This is odd, because the NetBSD rarpd uses htons(ETHERTYPE_REVARP). (Praise be to tcpdump: I would never have been able to track this silliness down without it.) - The update_arptab() function has been castrated. It depends on SIOCSARP which has been deprecated in 4.4BSD. The NetBSD people don't seem to be using this function either. It wouldn't be too hard to replace this with equivalent code from arp.c, but it might not be necessary. - I put together an ether_ntohost() support function that allows both local (/etc/ethers) and NIS lookups. This stuff should go in libc at some point, but nothing else seems to need it for now, so it can wait a while. As you may have guessed, you need to have the Berkeley Packet Filter in your kernel in order to use this program. The good news is that together with the recently added bootparamd, you can use finally use a FreeBSD box to boot Sun boxes over the network. (This was my whole motivation for getting this stuff to work: I have this one subnet that has a whole bunch of Sun3 X-terminals on it with only two Sun4 workstations, both of which are locked in peoples' offices. If those two machines crash (and they do every so often) then none of the X-terms will boot. Now I can use a spare PC that I have as a boot server. :)
1995-03-02 06:41:40 +00:00
}
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
/*
* A faster replacement for inet_ntoa().
*/
char *
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
intoa(u_long addr)
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
{
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
char *cp;
u_int byte;
int n;
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
static char buf[sizeof(".xxx.xxx.xxx.xxx")];
cp = &buf[sizeof buf];
*--cp = '\0';
n = 4;
do {
byte = addr & 0xff;
*--cp = byte % 10 + '0';
byte /= 10;
if (byte > 0) {
*--cp = byte % 10 + '0';
byte /= 10;
if (byte > 0)
*--cp = byte + '0';
}
*--cp = '.';
addr >>= 8;
} while (--n > 0);
return cp + 1;
}
char *
WARNS=2 cleanup, ANSIfication, manpage mdoc(7) cleanup. Once again, as explained in my messages to -audit, the ANSIfication comes as part of the preparation to add a new -d command-line flag to send output to stdout/stderr. That commit will come in a week, pending any further comments/objections. For those who have missed the -audit mails, it's at http://people.FreeBSD.org/~roam/bsd/rarpd/usr.sbin-rarpd-d.patch Asbestos suit: on ;) Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-11 09:29:34 +00:00
eatoa(u_char *ea)
Fix up new rarpd. This includes the following changes: - Support for poking ARP entries into the local table is now built in, so the arptab.c module I hacked together is no longer needed. - rarp_process() and rarp_reply() now accept a len argument which is passed down from rarp_loop() which tells rarp_reply() exactly how long the original RARP frame was. (Usually, it's 60 bytes, which is the minimum.) Previously, the length was calculated using the sum of sizeof(struct ether_header) + sizeof(struct ether_arp) (plus the ethernet frame header, I think). The result was a total packet length of 42 bytes. Now, rarp_reply() sends out packets that are the same size as those it recieves (60 bytes). This agrees with the behavior of rarpd on SunOS (as observed with tcpdump). The unused extra bytes are zeroed.
1996-11-18 22:07:41 +00:00
{
static char buf[sizeof("xx:xx:xx:xx:xx:xx")];
(void)sprintf(buf, "%x:%x:%x:%x:%x:%x",
ea[0], ea[1], ea[2], ea[3], ea[4], ea[5]);
return (buf);
}
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
void
logmsg(int pri, const char *fmt, ...)
{
va_list v;
FILE *fp;
char *newfmt;
va_start(v, fmt);
if (dflag) {
if (pri == LOG_ERR)
fp = stderr;
else
fp = stdout;
if (expand_syslog_m(fmt, &newfmt) == -1) {
vfprintf(fp, fmt, v);
} else {
vfprintf(fp, newfmt, v);
free(newfmt);
}
fputs("\n", fp);
fflush(fp);
} else {
vsyslog(pri, fmt, v);
}
va_end(v);
}
int
expand_syslog_m(const char *fmt, char **newfmt) {
const char *str, *m;
char *p, *np;
p = strdup("");
str = fmt;
while ((m = strstr(str, "%m")) != NULL) {
make it compile on alpha again
2001-07-05 16:53:49 +00:00
asprintf(&np, "%s%.*s%s", p, (int)(m - str),
str, strerror(errno));
Add a -d command-line option; when used in conjunction with -f, rarpd sends error messages to stderr, normal output to stdout, instead of logging everything via syslog. Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how to disable the check for one single line in the source :( Reviewed by: dd, silence on -audit MFC after: 1 month
2001-06-18 06:48:33 +00:00
free(p);
if (np == NULL) {
errno = ENOMEM;
return (-1);
}
p = np;
str = m + 2;
}
if (*str != '\0') {
asprintf(&np, "%s%s", p, str);
free(p);
if (np == NULL) {
errno = ENOMEM;
return (-1);
}
p = np;
}
*newfmt = p;
return (0);
}
Reference in New Issue Copy Permalink