April 24, 2008
View c7337a2747

added CRLF ping/pong keepalives aka SIP outbound

Alfred E. Heggestad authored on 24/04/2008 13:51:26
February 5, 2008
View 7498b4dc09

- avoid extra read syscalls (detect short reads) in poll_et or sigio_rt case (should improve performace) - if POLLRDHUP or EPOLLRDHUP are supported (linux >= 2.6.17) use them and avoid an extra syscall on EOF - on write error try to see if there's still some data in the socket read buffer and try to process it first (if there's no more data do a quick connection destroy)

Andrei Pelinescu-Onciul authored on 05/02/2008 21:47:29
January 12, 2008
View 6a8978e53c

- tcp optimization: in tcp_reader try reading from a fd immediately after receiving it (since it receives it only if there's data waiting on it => avoids a few sys calls before starting readingi or for a bad connect attempt)

Andrei Pelinescu-Onciul authored on 12/01/2008 15:17:40
January 8, 2008
View 057063e685

- tcp: fix local timer intialization (proper prev_ticks init.)

Andrei Pelinescu-Onciul authored on 08/01/2008 13:11:30
December 22, 2007
View b4fa727e63

- fix: tcp fd cache: don't cache own fd in a tcp_reader - various cleanups/better error messages

Andrei Pelinescu-Onciul authored on 22/12/2007 08:11:14
December 21, 2007
View 316130a735

- fixes: - on io_watch_add overwrite error don't delete the previous fd hash entry + more error debugging info - return proper len on pending connect (instead of -1) - tcp_reader: always check if a connection was marked as bad and if so release it immediately + error checks for io_watch_*

Andrei Pelinescu-Onciul authored on 21/12/2007 23:27:13
December 11, 2007
View 8b0472d783

- local_timer cleanup (always use local_timer*, don't mix them with timer_*) - tcp fix: clear the write watch flag after empting the write buffers

Andrei Pelinescu-Onciul authored on 11/12/2007 20:11:01
December 5, 2007
View 9188021ac6

Initialize and destory the configuration framework in main.c, and update the local config within the child processes at the beginning of each loop.

Miklos Tirpak authored on 05/12/2007 15:51:24
November 29, 2007
View a0553f4e1e

- io_wait support for write - io_wait: added io_watch_chg(..) - updated tcp code to the io_wait api changes

Andrei Pelinescu-Onciul authored on 29/11/2007 21:01:45
November 27, 2007
View 5c5cd7369d

- for tcp read processes, reuse the read fd for sending - keep a tcp send fd cache (experimental) (performance improvement)

Andrei Pelinescu-Onciul authored on 27/11/2007 21:05:32
View ccb7fda23b

- switched to much better tcp timers (performance increase especially with tesn of thousands of active tcp connections)

Andrei Pelinescu-Onciul authored on 27/11/2007 07:23:27
July 4, 2007
View 3e99928136

- added register_fds(no), used to register the maximum numbers of fds opened by a module (on a per process basis). Needed for code that needs to know the maximum possible open fd number (like the tcp code). - added get_max_open_fds()

Andrei Pelinescu-Onciul authored on 04/07/2007 17:27:33
February 21, 2007
View 49d12f5fef

- fixed get_ticks* usage: only get_ticks_raw() must be used and everything must be converted to ticks. Thanks to Alfred E. Heggestad" <aeh@db.org> for spotting this.

Andrei Pelinescu-Onciul authored on 21/02/2007 01:42:25
February 20, 2007
View 1eee4c4ddb

- fixed wrong expire time calculation (not correct when tick oveflow, an timeout == some value before the overflow). This should fix the connection lives forever bug reported by Alfred E. Heggestad <aeh@db.org>.

Andrei Pelinescu-Onciul authored on 20/02/2007 18:20:15
February 16, 2007
View 301fa31498

- minor: nicer error handling

Andrei Pelinescu-Onciul authored on 16/02/2007 21:34:13