Commit graph

1537 commits

Author SHA1 Message Date
Martin Mares
ea0ac8f69a Move CLI socket to the newly created CLI's pool.
(thanks to Andreas for the original idea)
2004-05-31 22:00:18 +00:00
Martin Mares
38a608c55a Rewritten the I/O loop. All socket operations are now safe, meaning that
you can delete the socket from anywhere in the hooks and nothing should break.
Also, the receive/transmit buffers are now regular xmalloc()'ed buffers,
not separate resources which would need shuffling around between pools.

sk_close() is gone, use rfree() instead.
2004-05-31 21:48:19 +00:00
Martin Mares
206f59dfa8 Added UNUSED. 2004-05-31 21:02:09 +00:00
Martin Mares
6a57bb3110 Killed a couple of unused variables.
We really should compile with warnings enabled.
2004-05-31 20:57:38 +00:00
Martin Mares
d83faf8dc4 static declarations don't belong to includes. (And most of them were
redundant anyway.)
2004-05-31 20:53:22 +00:00
Martin Mares
7deffd845a Need <termios.h> for tcdrain(). 2004-05-31 20:51:45 +00:00
Martin Mares
9f387e11a3 Make the check for <linux/rtnetlink.h> work with recent libc's. 2004-05-31 20:49:11 +00:00
Martin Mares
0757bcb728 One space more. 2004-05-31 20:35:19 +00:00
Martin Mares
2cc37815ae Added rmove() (by Andreas, tweaked by me). 2004-05-31 18:47:19 +00:00
Martin Mares
0077aab4f9 The code was broken for external /29 to /32 routes. Assuming that you
have one machine publishing a route to 10.1.1.3/32 and another one
publishing a route to 10.1.1.4/32. If the first machine went down the
route to 10.1.1.4/32 was wrongly killed by the old code, leading either
to missing routes or worse to bug()s like "Router parent does not have
next hop" or just segfaults. The patch fixes this but in the long term a
redesign is required here. Note that the patch doesn't worse the
situation, instead it prevents the problems stated. The redesign is
required to handle multiple routes to small subnets properly.

(by Andreas)

Feela, I think that this is at least a good temporary fix, but it's
of course up to you to decide.
2004-05-31 18:16:42 +00:00
Martin Mares
4ef3dccfa1 The OSPF authentication type was sent in host byte order instead of of
network byte order thus breaking interoperability with other routing
daemons on litte endian machines. The patch fixes this but note that
this breaks compatability with older bird installations using OSPF and
password authentication

(Andreas)
2004-05-31 18:13:14 +00:00
Martin Mares
3cb96cd343 The initial sequence number for RIP md5 authentication was always zero.
Bad as when bird e.g. was running for two weeks and then restarted it
would take another two weeks until the peers of this router would accept
data again from this router, as the sequence number would be too low.
Changed to use the the current system time as the starting sequence
number which is a more sane start value.

(by Andreas, cleaned up by me)
2004-05-31 18:11:16 +00:00
Martin Mares
277a34eff1 Small correction to va_start/va_end in cli_printf (va_end was missing).
(Andreas)
2004-05-31 18:08:50 +00:00
Martin Mares
5f2a6a9ff3 Fix handling on full pipe to client in bird. Prevent packet overflows
for even only medium sized route table output. Fix a strange garbled
output problem in the client. The latter seems to be caused by some
library doing tcflush while there is still command output pending. So
the best fix here is to do fflush and then tcdrain. Note that this
problem occurs only under certain load situations and is not too easy to
reproduce.

(by Andreas)
2004-05-31 17:55:30 +00:00
Martin Mares
a4ffe2739d Bird's control socket should be in /var/run and the convention for
--localstatedir is /var. The control socket pathname creation is thus
corrected here.

(Andreas)
2004-05-31 17:53:02 +00:00
Martin Mares
03e3d184b2 Fix bison input for current build tools, otherwise bison or the compiler
will abort the build.

(by Andreas)
2004-05-31 17:44:39 +00:00
Martin Mares
4a02013767 Make RIP nolisten mode actually work. The socket is required for
sending, the received data has to be discarded instead.

(patch by Andreas Steinmetz modified by me)
2004-05-31 17:42:38 +00:00
Ondrej Filip
1bd897dd33 Changed of comments. 2004-05-31 17:27:21 +00:00
Ondrej Filip
10af3676ea Grrr, committing too fast.
#include "alloca.h" -> #include "lib/alloca.h"
2004-05-31 17:16:47 +00:00
Ondrej Filip
7dbd4fd332 alloca.h added 2004-05-31 17:13:58 +00:00
Ondrej Filip
0e6eef620d Use #include "alloca.h" 2004-05-31 17:07:05 +00:00
Ondrej Filip
c222500d8e Previous change was mistake. 2004-05-31 17:00:22 +00:00
Ondrej Filip
6f18235aad Useless include deleted. 2004-05-31 16:42:12 +00:00
Martin Mares
f54fa9e15d Updated copyright notices.
Also testing whether syncmail works.
2004-05-31 16:10:01 +00:00
Ondrej Filip
73219ecfec Delete autom4te.cache in 'make distclean'. 2004-05-31 15:13:56 +00:00
Ondrej Filip
6de6292312 Better readline checking. 2004-05-31 15:08:29 +00:00
Ondrej Filip
012279f395 Some include added. 2004-05-31 13:59:03 +00:00
Ondrej Filip
d93fb7e6b9 #ifndef ALIGN - it is defined on *BSD 2004-05-31 13:58:38 +00:00
Ondrej Filip
a60277b999 Added RTD_NONE /* Just for internal use */ 2004-05-31 13:35:06 +00:00
Ondrej Filip
0c745adc80 #ifdef ALLOCA_H 2004-05-31 13:34:20 +00:00
Ondrej Filip
bd62eeca27 Small change to compile client on FreeBSD. 2004-05-31 13:32:58 +00:00
Ondrej Filip
b1a1fabac7 *BSD port added. (Tested on FreeBSD and NetBSD) 2004-05-31 13:25:00 +00:00
Ondrej Filip
781aa475aa Minor bug fix in neighbor state machine. 2004-05-31 13:22:49 +00:00
Ondrej Filip
de259a41e3 Minor bug in configuration. 2004-05-31 10:38:44 +00:00
Ondrej Filip
7cb37e6fd2 Added some more test (alloca.h, sa_len) etc.
Add AC_OUTPUT before AC_OUTPUT_COMMANDS
2004-05-31 10:26:18 +00:00
Ondrej Filip
caeb02ea19 Bug in DBDES receiving fixed. 2004-05-19 12:30:58 +00:00
Martin Mares
502ded5215 Fix reporting of RIP socket errors. Thanks to Eric Leblond for the patch. 2003-12-06 16:41:11 +00:00
Ondrej Filip
37299f1e44 OSPF was not able to be built stand-alone. 2003-09-30 17:05:55 +00:00
Ondrej Filip
00bd27a1cc Endianity problem fixed. Thanx to Sörös József 2003-09-14 13:41:24 +00:00
Ondrej Filip
35a86ceb40 This prevents infinite loop when bird has more that 60 neighbors. Thanks to Rani Assaf 2003-09-03 17:31:23 +00:00
Ondrej Filip
bf135bcb1f Prepared for release 1.0.8. 2003-09-01 08:46:07 +00:00
Ondrej Filip
c11007bc42 Endian-related bug fixes sent by Krzysztof Szuster <kszuster@polbox.com> 2003-08-26 10:41:02 +00:00
Ondrej Filip
1d1a3c1c2a Minor OSPF changes for faster startup. 2003-08-23 10:47:46 +00:00
Ondrej Filip
baa5dd6ccc Many spelling mistakes fixed.
Thanks you Harry Moyes.
2003-08-23 10:42:41 +00:00
Martin Mares
c197d44e17 This probably IS the memory leak we're looking for. Alien routes weren't
correctly disposed of.
2003-08-13 22:07:55 +00:00
Martin Mares
13b75bacf7 protocol->import_control() could potentially call rte_cow() as well.
AFAIK it doesn't happen in any of our protocol, but better be sure.
2003-08-13 20:04:39 +00:00
Martin Mares
2adab6ae9c This was a potential memory leak, but not the one Feela observed.
This one could happen when an import filter of some protocol modified
the rte (so that it would be rte_cow()ed) and later rejected it.
2003-08-13 19:31:22 +00:00
Martin Mares
b77834b3f4 Slept for a year :-) 2003-04-06 21:36:35 +00:00
Martin Mares
c153ee9137 Nobody is perfect. Me twice :) 2003-04-06 19:52:22 +00:00
Martin Mares
8edd56bded Oops, forgot to change some paths. 2003-04-06 19:49:17 +00:00