Commit Graph

69 Commits

Author SHA1 Message Date
William Pitcock af6f5d479f GNUTLS backend. Untested. 2008-06-10 20:33:15 -05:00
William Pitcock dd91040957 Run autoreconf. 2008-06-10 13:45:43 -05:00
William Pitcock 5d92990907 libratbox: Add checks for GnuTLS support. 2008-06-10 13:45:15 -05:00
Jilles Tjoelker 6315153b9c libratbox: Remove static inline symbols from export-syms.txt. 2008-05-17 13:26:42 +02:00
Jilles Tjoelker e0bd36a502 Fix a mistake in kqueue 'overflow' handling.
If there is no space in the output buffer to report an
error adding to the kqueue, kevent(2) will abort and
return the error in errno (I was correct that it does
not tell you where it failed). So do not abort the loop
if kevent(2) fails and do not log (expected) EBADF.
2008-05-14 19:56:41 +02:00
Jilles Tjoelker 9b4f3c47e1 Save connect errno so that we get a correct connect failure reason
libratbox r25358 (androsyn)
2008-05-13 20:21:12 +02:00
Jilles Tjoelker b68b0b2cde Fix a possible crash with SSL connections closing early.
libratbox r25356 (jilles)
2008-05-12 19:54:24 +02:00
Jilles Tjoelker 98686f18e9 Fix kqueue sometimes dropping updates.
(ircd wouldn't read or write anymore to certain clients)
This happens because kqueue.c will often try to add
already closed file descriptors to the kqueue. The kernel
tries to report bad file descriptors in the eventlist; if
the eventlist has no space, processing of the changelist
is silently halted.
The fix:
1. allocate two kqlst things, one for what kqlst currently does
and one as output buffer
this ensures the kevent(2) call in rb_select_kqueue() never
drops updates
2. replace the kevent(2) call in kq_update_events() by a loop
that processes the updates one at a time
that doesn't happen much, and it's the only way to be sure
without also getting events out of the queue we cannot process
at that time
libratbox r25354 (jilles)
2008-05-12 18:54:20 +02:00
Jilles Tjoelker 79c4d759a9 patricia: remove ugly abuse of K&R style function declaration
It left the argument types unspecified in a function
pointer, then called it using different numbers of
arguments.
libratbox r25229
2008-04-13 18:20:18 +02:00
Jilles Tjoelker dcb90e0d86 Remove linebuf's per-line flushing flag, as it's per-head state.
In rare cases, this sharing caused the ircd to skip
part of outgoing traffic, e.g. appearing as "not enough
parameters" errors on the other side.
The purpose of this flag can be fulfilled by the writeofs
in the bufhead.
libratbox r25227
2008-04-13 00:44:21 +02:00
Jilles Tjoelker ec9738cbd8 read/write return type should be ssize_t, not int or size_t.
The writev emulation used size_t, which is unsigned,
preventing negative values to be seen.
libratbox r25225
2008-04-12 16:43:12 +02:00
Jilles Tjoelker c50bb62ccd kqueue: also use EV_ONESHOT for read events
This makes the kernel's state agree with our handler pointer.
SSL may need to suspend selecting for reading to write
something, e.g. with renegotiation.
libratbox r25223
2008-04-11 11:08:37 +02:00
Valery Yatsko 8f40f4bb7f Importing r25217, r25219 and r25221 from ratbox3 2008-04-10 20:37:42 +04:00
Valery Yatsko cee7d8576d Importing changes from ircd-ratbox revision r25203, this fixes libratbox/src/openssl.c a bit 2008-04-06 18:28:56 +04:00
Jilles Tjoelker b5d9e56cfb Fix SET_SS_LEN so that it also works correctly outside libratbox.
Use RB_SOCKADDR_HAS_SA_LEN instead of SOCKADDR_IN_HAS_LEN
which is only defined inside libratbox.
This fixes creating own sockaddr structs in other programs
using SET_SS_LEN, then passing them to libratbox, on
4.4BSD based systems.
2008-04-05 21:08:48 +02:00
Valery Yatsko 6e9b4415cc ok, trying to work on blockheap's stuff 2008-04-02 04:28:05 +04:00
William Pitcock 1ac6502183 Make this link without OpenSSL. This patch should go upstream. 2008-04-01 14:57:52 -05:00
William Pitcock 56efe9745d Run autoreconf. 2008-04-01 11:53:46 -05:00
William Pitcock b57f37fb6a Pull in libratbox from ircd-ratbox-3.0.0beta1 and integrate into the build system. 2008-04-01 11:52:26 -05:00