January 3, 2015
View 2266bb9

Core Update of doxygen, removal of history, changing "ser" to "kamailio"

Olle E. Johansson authored on 03/01/2015 14:03:24
April 2, 2012
View abb01fb

core: Added void to declarations of functions with empty argument list

Marius Zbihlei authored on 02/04/2012 14:19:17
January 6, 2012
View c2429a5

core: new cfg parameter fork_delay

Daniel-Constantin Mierla authored on 06/01/2012 11:06:02
June 29, 2011
View 23bd471

core: fix two more 'no real prototype' warnings, fix one wrong caller in timer mod

Henning Westerholt authored on 29/06/2011 20:30:39
August 19, 2010
View 83e91df

core: daemon status/pipe fixes & interface changes

Andrei Pelinescu-Onciul authored on 19/08/2010 18:03:54
March 13, 2009
View 33bfeb9

The fixup function prototypes of the config variables have been extended with the group name. All the fixup functions have been updated.

Miklos Tirpak authored on 13/03/2009 13:59:28
April 4, 2008
View b4c9f98

Two configuration variables are introduced that can help troubleshooting memory leaks: mem_dump_pkg, and mem_dump_shm Useage:

Miklos Tirpak authored on 04/04/2008 08:40:53
December 5, 2007
View 63b50c5

- close all the unneeded fds in child processes (=> significantly less fds used)

Andrei Pelinescu-Onciul authored on 05/12/2007 20:26:58
July 7, 2007
View faa66b4

- added missing register_fds & get_max_open_fds() declarations

Andrei Pelinescu-Onciul authored on 07/07/2007 00:33:02
September 20, 2006
View 4fd7c9f

- added ser profiling patch from Hendrik Scholz <hendrik.scholz@freenet-ag.de>, adapted to the new fork functions (for more info see http://lists.iptel.org/pipermail/serdev/2006-July/007627.html)

Andrei Pelinescu-Onciul authored on 20/09/2006 10:28:42
View 2254e60

- allow compilation without USE_TCP option

Hendrik Scholz authored on 20/09/2006 09:22:48
September 19, 2006
View 37209e1

- applied patch from Dragos Vingarzan <vingarzan@fokus.fraunhofer.de> which moves all the forking part into 2 functions in pt.c (fork_process and fork_tcp_process). - added PROC_NOCHLDINIT rank value for Dragos's fork_process (if this value is used as child_id/rank_value the mod_child functions will not be called) - added register_procs(processes_no), used from mod_init when a module knows that it will fork some children (replaces the old process_count++ / the patch's estimated_process_count++) - added get_max_procs(): returns the maximum (estimated) number of processes

Andrei Pelinescu-Onciul authored on 19/09/2006 16:13:27
December 13, 2005
View 55a4526

- Modules can properly register processes in SER process table - process_count function has been replaced by process_count variable - modules can increase the value of process_count in mod_init, this will be used to allocate process table that will be big enough - modules can create processes in child_init with rank PROC_MAIN - the index into process table is in last_process, this is where information about the new process should be recorded - do not forget to reset is_main in the new process otherwise it will not be terminated properly - modules should increate last_process before fork - the new child should set process_no = last_process

Jan Janak authored on 13/12/2005 12:10:15
December 5, 2005
View af93cbd

New timer interface: - allows adding timers dynamically - supports one shot and periodic timers - precise - allows timeouts < 0.5 s (depends on the config., by default min. timeout = 62.5 ms) - based on a 3 level hierarchical timing wheel - very fast - low overhead for timer functions (most of the time, the excetion is timer_del which take can take quite a long time in special situations) - supports "slow" timers (compile option, by default on), "slow" timer = a timer handler which is declared as possible taking a long timer to execute. - backward compatible: old timer functions are preserverd => no changes needed for the existing code (although migration to the new interface is recommended) For more information read doc/timers.txt. tm timer update will follow shortly (this week).

Andrei Pelinescu-Onciul authored on 05/12/2005 18:29:30
November 26, 2004
View f1aaf28

- process_count was wrong when tcp was disabled

Andrei Pelinescu-Onciul authored on 26/11/2004 16:27:22