segfault with exabgp and bird 1.4.2
Hi, I have been running a stable mesh with a BIRD peer in a mesh for a few weeks now. Trying to connect with ExaBGP makes BIRD segfault. How can I debug this further? Here is the TCP stream: https://spodder.com/~yeled/bird.pcap It seems innocuous to me. C. ~% ulimit -c 100000 ~% file ./bird ./bird: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), for GNU/Linux 2.6.8, dynamically linked (uses shared libs), not stripped ~% sudo ./bird bird: Hello or poll interval less that 5 makes cryptographic authenication prone to replay attacks ~% sudo ./birdc BIRD 1.4.2 ready. Connection closed by server. ~% sudo gdb bird core GNU gdb 6.8-debian Copyright (C) 2008 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "x86_64-linux-gnu"... warning: Can't read pathname for load map: Input/output error. Reading symbols from /lib/librt.so.1...done. Loaded symbols for /lib/librt.so.1 Reading symbols from /lib/libpthread.so.0...done. Loaded symbols for /lib/libpthread.so.0 Reading symbols from /lib/libc.so.6...done. Loaded symbols for /lib/libc.so.6 Reading symbols from /lib/ld-linux-x86-64.so.2...done. Loaded symbols for /lib64/ld-linux-x86-64.so.2 Core was generated by `./bird'. Program terminated with signal 11, Segmentation fault. [New process 11480] #0 bgp_rx (sk=0x6aa930, size=<value optimized out>) at ../../../proto/bgp/packets.c:1030 1030 ../../../proto/bgp/packets.c: No such file or directory. in ../../../proto/bgp/packets.c (gdb) bt #0 bgp_rx (sk=0x6aa930, size=<value optimized out>) at ../../../proto/bgp/packets.c:1030 #1 0x00000000004396f3 in sk_read (s=0x6aa930) at io.c:1602 #2 0x000000000043ab03 in io_loop () at io.c:1843 #3 0x000000000043d7a1 in main (argc=<value optimized out>, argv=<value optimized out>) at main.c:820 (gdb)
On Sat, May 03, 2014 at 02:37:18PM +0100, Charlie Allom wrote:
Hi,
I have been running a stable mesh with a BIRD peer in a mesh for a few weeks now.
Trying to connect with ExaBGP makes BIRD segfault. How can I debug this further?
Could you send me core dump and your ExaBGP config? You could also try 1.4.3 (1.4.2 has known bug, but this seems like some different problem). -- Elen sila lumenn' omentielvo Ondrej 'SanTiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
On Sat, May 03, 2014 at 05:42:24PM +0200, Ondrej Zajicek wrote:
On Sat, May 03, 2014 at 02:37:18PM +0100, Charlie Allom wrote:
Hi,
I have been running a stable mesh with a BIRD peer in a mesh for a few weeks now.
Trying to connect with ExaBGP makes BIRD segfault. How can I debug this further?
Could you send me core dump and your ExaBGP config?
And your ExaBGP version. -- Elen sila lumenn' omentielvo Ondrej 'SanTiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
ExaBGP fails with either 3.2.19 or 3.3.2 (HEAD) I will send you the core and the config privately, if you don't mind. On Sat, May 3, 2014 at 5:00 PM, Ondrej Zajicek <santiago@crfreenet.org> wrote:
On Sat, May 03, 2014 at 05:42:24PM +0200, Ondrej Zajicek wrote:
On Sat, May 03, 2014 at 02:37:18PM +0100, Charlie Allom wrote:
Hi,
I have been running a stable mesh with a BIRD peer in a mesh for a few weeks now.
Trying to connect with ExaBGP makes BIRD segfault. How can I debug this further?
Could you send me core dump and your ExaBGP config?
And your ExaBGP version.
-- Elen sila lumenn' omentielvo
Ondrej 'SanTiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux)
iEYEARECAAYFAlNlEp8ACgkQw1GB2RHercPuugCdExG7zXtftI5LY1ubmo+BGbSx qxUAoIDSxxHs4m4RSHigaAfw/bI6UZir =MOtF -----END PGP SIGNATURE-----
-- charlie@yelp.com +44.207.347.1682
hmmm, i was sure I checked this, but it seems a false alarm. ExaBGP HEAD (v3.3.2) seems to not crash BIRD. sorry for the noise. On Sat, May 3, 2014 at 5:00 PM, Ondrej Zajicek <santiago@crfreenet.org> wrote:
On Sat, May 03, 2014 at 05:42:24PM +0200, Ondrej Zajicek wrote:
On Sat, May 03, 2014 at 02:37:18PM +0100, Charlie Allom wrote:
Hi,
I have been running a stable mesh with a BIRD peer in a mesh for a few weeks now.
Trying to connect with ExaBGP makes BIRD segfault. How can I debug this further?
Could you send me core dump and your ExaBGP config?
And your ExaBGP version.
-- Elen sila lumenn' omentielvo
Ondrej 'SanTiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux)
iEYEARECAAYFAlNlEp8ACgkQw1GB2RHercPuugCdExG7zXtftI5LY1ubmo+BGbSx qxUAoIDSxxHs4m4RSHigaAfw/bI6UZir =MOtF -----END PGP SIGNATURE-----
-- charlie@yelp.com +44.207.347.1682
On Sat, 03 May 2014, Charlie Allom wrote:
hmmm, i was sure I checked this, but it seems a false alarm.
If it managed to crash bird 1.4.2, bird is buggy regardless of whatever ExaBGP fed it. So, it doesn't matter if recent ExaBGB doesn't trigger the bird bug anymore, what matters is whether latest bird can still be crashed by old ExaBGP :-) -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh
fair enough, I will follow up with Ondrej with the coredump and binary and pcap as I can replicate it :) On Sat, May 3, 2014 at 8:20 PM, Henrique de Moraes Holschuh <hmh@hmh.eng.br> wrote:
On Sat, 03 May 2014, Charlie Allom wrote:
hmmm, i was sure I checked this, but it seems a false alarm.
If it managed to crash bird 1.4.2, bird is buggy regardless of whatever ExaBGP fed it.
So, it doesn't matter if recent ExaBGB doesn't trigger the bird bug anymore, what matters is whether latest bird can still be crashed by old ExaBGP :-)
-- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh
-- charlie@yelp.com +44.207.347.1682
On Sat, May 3, 2014 at 8:20 PM, Henrique de Moraes Holschuh
If it managed to crash bird 1.4.2, bird is buggy regardless of whatever ExaBGP fed it.
Exactly, it shouldn't crash for any input. On Sat, May 03, 2014 at 08:59:10PM +0100, Charlie Allom wrote:
fair enough, I will follow up with Ondrej with the coredump and binary and pcap as I can replicate it :)
Not needed. I already managed to replicate it with ExaBGP 3.2.19 . -- Elen sila lumenn' omentielvo Ondrej 'SanTiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
what do i win? :) Thanks Ondrej! On Sat, May 3, 2014 at 9:48 PM, Ondrej Zajicek <santiago@crfreenet.org> wrote:
On Sat, May 3, 2014 at 8:20 PM, Henrique de Moraes Holschuh
If it managed to crash bird 1.4.2, bird is buggy regardless of whatever ExaBGP fed it.
Exactly, it shouldn't crash for any input.
On Sat, May 03, 2014 at 08:59:10PM +0100, Charlie Allom wrote:
fair enough, I will follow up with Ondrej with the coredump and binary and pcap as I can replicate it :)
Not needed. I already managed to replicate it with ExaBGP 3.2.19 .
-- Elen sila lumenn' omentielvo
Ondrej 'SanTiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux)
iEYEARECAAYFAlNlVgsACgkQw1GB2RHercNA2gCfbooQUbOV4CCHx0+oKU69BQwn CdcAn13bU8W/sQl1S5jSTxAfvtPhinEH =Yhaw -----END PGP SIGNATURE-----
-- charlie@yelp.com +44.207.347.1682
On Sat, May 03, 2014 at 09:19:22PM +0100, Charlie Allom wrote:
what do i win? :)
A bugfix patch (attached) ;-) -- Elen sila lumenn' omentielvo Ondrej 'SanTiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
participants (3)
-
Charlie Allom -
Henrique de Moraes Holschuh -
Ondrej Zajicek