[Asterisk-code-review] chan sip: Don't unload module while TCP/TLS threads are stil... (asterisk[master])
Corey Farrell
asteriskteam at digium.com
Wed Dec 7 13:07:09 CST 2016
Corey Farrell has posted comments on this change. ( https://gerrit.asterisk.org/4536 )
Change subject: chan_sip: Don't unload module while TCP/TLS threads are still running
......................................................................
Patch Set 3: Code-Review-1
(3 comments)
I have not commented on all the issues with this patch. Please allow me to post a patch for this, I have a pretty good idea of how chan_sip unload can / can't be reordered. I have a patch nearly completed but need a little more time to scrutinize it before I can post.
https://gerrit.asterisk.org/#/c/4536/3/channels/chan_sip.c
File channels/chan_sip.c:
Line 1783: if (event_state_compositors[i].compositor) {
The if statement is not needed, it is safe to call ao2_cleanup on a NULL value (so setting NULL is all that is needed.
Line 35200: if (!threadt) {
Unclear why this conditional is needed, load_module should never be run unless unload_module was completely successful (return 0).
I'll have to double-check the code in loader.c, but if this is needed I'm inclined to say that's a bug in the module loader.
PS3, Line 35533: default_tls_cfg.certfile = ast_strdup("");
: default_tls_cfg.pvtfile = ast_strdup("");
: default_tls_cfg.cipher = ast_strdup("");
: default_tls_cfg.cafile = ast_strdup("");
: default_tls_cfg.capath = ast_strdup("");
No can do. Each of these are memory leaks.
--
To view, visit https://gerrit.asterisk.org/4536
To unsubscribe, visit https://gerrit.asterisk.org/settings
Gerrit-MessageType: comment
Gerrit-Change-Id: I23a8abf9e8f73b57ff7ada61984b8506aea6cbde
Gerrit-PatchSet: 3
Gerrit-Project: asterisk
Gerrit-Branch: master
Gerrit-Owner: Michael Kuron <m.kuron at gmx.de>
Gerrit-Reviewer: Anonymous Coward #1000019
Gerrit-Reviewer: Corey Farrell <git at cfware.com>
Gerrit-Reviewer: Michael Kuron <m.kuron at gmx.de>
Gerrit-HasComments: Yes
More information about the asterisk-code-review
mailing list