June 12, 2007
View fbe119140b

- added sf_malloc: a multi-process safe (internal locking), multi-pool f_malloc version. Should perform better on lots of CPU once properly tunned (experimental for now) - added ll_malloc: same as above but most operation are lockless (except for "big" fragment handling). For now needs tunning and it still keeps too many debugging statistics. (experimental for now)

Andrei Pelinescu-Onciul authored on 12/06/2007 18:07:12
June 7, 2007
View 8390f72218

- added the new PROC_INIT rank and the init_child(PROC_INIT) calls (see doc/modules_init.txt for more info) - introduced new options for locking pages in memory, pre-faulting the shared memory pages and real time mode

Andrei Pelinescu-Onciul authored on 07/06/2007 21:40:14
May 26, 2007
View 17cde66541

- darwin sigwait() bug workaround: on darwin sigwait is buggy: it will cause extreme slow down on signal delivery for the signals it doesn't wait on (I've measured a 36s delay on a 1.5Ghz powerbook). The workaround is: sigwait() on all the signals we are interested in ser and manually call the master signal handler (sig_usr(sig_no)) if the signal!= slow timer signal.

Andrei Pelinescu-Onciul authored on 26/05/2007 11:12:29
May 14, 2007
View 8b8fc486bd

- added futex based locking, used by default on linux 2.5.70+ (to turn it off compile with make use_futex=no) - moved get_sys_ver from io_wait.c into ut.c

Andrei Pelinescu-Onciul authored on 14/05/2007 21:29:46
March 1, 2007
View f238de3d00

- sercmd ps udp receiver description fixed to be the same as in 0.9.6 (patch from Hendrik Scholz)

Andrei Pelinescu-Onciul authored on 01/03/2007 13:52:09
February 23, 2007
View b0123c365a

- relax db_init BUG check (allow db init from fixup functions)

Andrei Pelinescu-Onciul authored on 23/02/2007 21:03:38
February 10, 2007
View 6c53d41a09

- tls hooks support: special tls hooks added in core that allow a module or core part to provide a tls implementation. Support for the old in-core tls compile options is preserved (but one can compile with either tls hooks support or tls-in-core support, not with both in the same time). Changes were kept to a minimum. - Makefiles: make TLS=1 deprecated, use instead make TLS_HOOKS=1 or make CORE_TLS=1 (if in-core tls support is needed). Added TLS_EXTRA_LIBS (e.g. make CORE_TLS=1 TLS_EXTRA_LIBS="-lz -lkrb5").

Andrei Pelinescu-Onciul authored on 10/02/2007 18:52:48
January 24, 2007
View 8628de2247

- tls fix: set openssl locking functions (should be stable even under high load, lost of connections, and multi-cpu machines)

Andrei Pelinescu-Onciul authored on 24/01/2007 18:01:54
January 18, 2007
View d4fb00dc1d

- deadlock on no-detach (-DD) start-up error shutdown fixed (closes SER-210)

Andrei Pelinescu-Onciul authored on 18/01/2007 20:01:37
December 11, 2006
View ac34f9f4eb

- added special callbacks for non-sip msg (e.g http), They are called every time a message with the first line in sip/http format is received and the protocol != SIP. These callbacks are needed as a part of an xmlrpc bug fix.

Andrei Pelinescu-Onciul authored on 11/12/2006 15:47:32
November 28, 2006
View 279377197c

- random number are seeded in all the processes (both srand() and srandom())

Andrei Pelinescu-Onciul authored on 28/11/2006 19:04:54
November 4, 2006
View 5b84defdc6

- safer child exit on signal (s/exit()/_exit())

Andrei Pelinescu-Onciul authored on 04/11/2006 01:31:06
October 25, 2006
View f2e1aa5002

- init_childs(PROC_MAIN) moved before starting tcp_main to allow possible tcp usage from module started processes - fork_process() sanity checks & debugging: fail & log a bug if called with make_sock==1 from a process != main or if called after tcp was started; reset is_main for children - fork_tcp_process() sanity checks as above - init_pt(): initialize tcp comm. fds to -1, fix description writing for main/attendant.

Andrei Pelinescu-Onciul authored on 25/10/2006 18:44:36
View 96d091075b

- log messages in children signal handlers off if NO_SIG_DEBUG is defined (safer, but not default for now) - new config param.: exit_timeout (how much time ser will wait for its shutdown to complete, when it expires it will kill everything) - missing timeout added to SIGTERM triggered shutdown

Andrei Pelinescu-Onciul authored on 25/10/2006 16:44:21
October 13, 2006
View 96227c65df

- added STUN keep-alive functionality in accordance with draft-ietf-behave-rfc3489bis-04.txt

Vladimir Marek authored on 13/10/2006 13:08:42