| 00:14.20 | coreyfarrell | gtjoseph: sorry if I'm misunderstanding how stasis works, but I thought it was an asynchronous event queue, meaning it can get behind on dispatching events. |
| 00:14.49 | gtjoseph | true |
| 00:15.14 | gtjoseph | but i'm not understanding the context of your question. |
| 00:15.48 | coreyfarrell | related to the chan_sip patch |
| 00:16.13 | gtjoseph | could you be a little more specific? :) |
| 00:16.23 | gtjoseph | are you looking at the patch itself? |
| 00:16.54 | coreyfarrell | yes sorry.. |
| 00:17.09 | coreyfarrell | using peer in mwi_event_cb without a reference |
| 00:17.25 | coreyfarrell | let me look into what file said |
| 00:17.55 | gtjoseph | ah, so file and i determined that the only way for mwi_event_cb to be called is if there's a valid peer |
| 00:18.19 | gtjoseph | i put back the unsubscribe_and_join in destroy_mailbox |
| 00:19.02 | gtjoseph | so if there's a message, the peer can't have gone away. it MAY be in the process of it being cleaned up however |
| 00:19.12 | gtjoseph | but that's ok |
| 00:19.42 | gtjoseph | we don't increment the refcount because otehrwise the peer will NEVER go away |
| 00:20.16 | coreyfarrell | yup I think I'm getting it.. I just downloaded the patch so I can take a better look |
| 00:20.31 | gtjoseph | cool, definitely let me know |
| 00:20.33 | coreyfarrell | now that I realize the unsubscribe_and_join is still there protecting shutdown |
| 00:20.50 | file | and in the darkness... that which we thought was the problem... saves us |
| 00:21.01 | gtjoseph | :) |
| 00:22.00 | gtjoseph | the big key is that mwi_event_cb doesn't have to call sip_find_peer/build_peer again. |
| 00:33.59 | coreyfarrell | ok I understand what you mean with not needing the reference, agreed. |
| 00:34.26 | coreyfarrell | but.. do we need to lock peer in clear_peer_mailboxes to use AST_LIST_REMOVE_HEAD? |
| 00:34.53 | gtjoseph | looks |
| 00:34.54 | coreyfarrell | because sip_send_mwi_to_peer calls get_cached_mwi with peer locked, and that iterates the mailboxes. |
| 00:35.37 | coreyfarrell | but we probably can't call destroy_mailbox with the peer locked because that could deadlock stasis? |
| 00:35.50 | gtjoseph | yeah |
| 00:36.12 | gtjoseph | the destroy will cause mwi_event_cb to get called again |
| 00:36.24 | gtjoseph | i think |
| 00:36.41 | gtjoseph | to be honest, i've been staring at this code for 12.5 hours now. |
| 00:36.49 | coreyfarrell | ouch |
| 00:37.13 | gtjoseph | <PROTECTED> |
| 00:38.21 | coreyfarrell | ok I'll post to gerrit about clear_mailboxes. enjoy your night. |
| 05:07.00 | *** join/#asterisk-dev anthm (~anthm@freeswitch/developer/anthm) |
| 05:55.55 | *** join/#asterisk-dev bof22 (~Thunderbi@185.13.183.107) |
| 06:09.25 | *** join/#asterisk-dev WIMPy (~wimpy@85.183.95.26) |
| 06:51.14 | *** join/#asterisk-dev CELYA_ (~Thunderbi@LStLambert-657-1-90-73.w80-13.abo.wanadoo.fr) |
| 07:13.43 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 07:13.43 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 07:32.00 | *** join/#asterisk-dev pchero_work (~pchero@109.70.54.56) |
| 10:01.35 | *** join/#asterisk-dev hehol (~hehol@gatekeeper.loca.net) |
| 12:28.24 | *** join/#asterisk-dev acidfoo_ (~acidfoo@modemcable002.114-70-69.static.videotron.ca) |
| 12:58.53 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 12:58.53 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 13:09.24 | *** join/#asterisk-dev leedm777 (textual@nat/digium/x-ygrheswnvsrrdzto) |
| 13:30.02 | file | The future is friendly |
| 13:36.13 | *** join/#asterisk-dev acidfoo_ (~acidfoo@modemcable002.114-70-69.static.videotron.ca) |
| 13:51.17 | *** join/#asterisk-dev newtonr (RustyNewto@nat/digium/x-gbteknbionqymgdl) |
| 13:51.17 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 14:15.15 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-fqjzoroktzqzoauf) |
| 14:22.19 | gtjoseph | i wonder if we should add a flag to chan_sip/peer to indicate that a destruction is in progress |
| 14:43.51 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-gesfaebienwpasfl) |
| 14:56.59 | *** join/#asterisk-dev jameswf (uid27319@gateway/web/irccloud.com/x-jazghfcthtovuzyp) |
| 15:05.07 | coreyfarrell | gtjoseph: I had the same thought, I think the following would be valid: #define peer_in_destruction(peer) (ao2_ref(peer, 0) == 0) |
| 15:14.49 | *** join/#asterisk-dev leedm777 (~textual@50-81-28-152.client.mchsi.com) |
| 15:47.41 | gtjoseph | good call |
| 16:06.37 | gtjoseph | coreyfarrell: i did the list empty check because clear_peer_mailboxes is called by peer setup functions and didn't want to have to lock and unlock only to find the list empty. |
| 16:07.44 | gtjoseph | and i just saw those comments so i'll push up another version. |
| 16:08.49 | coreyfarrell | sorry which comments? are you going to remove the list empty check? |
| 16:09.52 | gtjoseph | i'd prefer not to because of where clear_mailboxes is called from |
| 16:10.48 | coreyfarrell | but reading the list without a lock isn't valid for other places it is called from |
| 16:11.19 | coreyfarrell | and during setup the lock isn't a big deal, if nothing else has a reference to the peer it won't block waiting for the lock |
| 16:12.42 | gtjoseph | i'll remove it |
| 16:16.46 | coreyfarrell | so the check for (regseconds > 0) was unintentional? |
| 16:17.17 | gtjoseph | it was part of an earlier attempt to figure out why rtpcachefrienbds=no wasn't working |
| 16:17.24 | gtjoseph | i just forgot to remove it |
| 16:18.15 | coreyfarrell | ah ok thanks |
| 16:39.32 | *** join/#asterisk-dev jkroon (~jkroon@197.96.224.13) |
| 17:25.55 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00c08cc1a9d6942398.ipv6.abo.wanadoo.fr) |
| 18:00.57 | *** join/#asterisk-dev jkroon (~jkroon@dustpuppy.is.co.za) |
| 18:39.05 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-82-81-175-197.red.bezeqint.net) |
| 18:39.05 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 19:06.49 | *** join/#asterisk-dev luckman212 (~luckman21@unaffiliated/luckman212) |
| 19:14.00 | *** join/#asterisk-dev Marquis42 (~Marquis42@d14-69-130-192.try.wideopenwest.com) |
| 19:24.30 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-82-81-175-197.red.bezeqint.net) |
| 19:24.30 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 21:37.54 | *** join/#asterisk-dev cresl1n (Adium@asterisk/libpri-and-libss7-expert/Cresl1n) |
| 21:37.54 | *** mode/#asterisk-dev [+o cresl1n] by ChanServ |
| 21:38.40 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00c08cc1a9d6942398.ipv6.abo.wanadoo.fr) |
| 21:38.40 | *** join/#asterisk-dev sh_smith (~smuxi@cpe-76-174-26-91.socal.res.rr.com) |
| 21:38.40 | *** join/#asterisk-dev danjenkins (danjenkins@gateway/shell/firrre/x-cmbjrcmkvsoingoq) |
| 21:40.58 | *** join/#asterisk-dev otherwiseguy (~otherwise@136.61.225.198) |
| 21:40.58 | *** join/#asterisk-dev gtjoseph (~gtjoseph@unaffiliated/gtj) |
| 21:40.58 | *** join/#asterisk-dev dfgg (damian@195-154-165-176.rev.poneytelecom.eu) |
| 21:40.58 | *** join/#asterisk-dev tm1000 (sid6728@gateway/web/irccloud.com/x-lmsevzpkyjdleeej) |
| 21:41.11 | *** join/#asterisk-dev WIMPy (~wimpy@85.183.95.26) |
| 21:41.11 | *** join/#asterisk-dev _longines (~longines@static.183.80.201.138.clients.your-server.de) |
| 21:41.11 | *** join/#asterisk-dev libardi (~kvirc@179.159.8.239) |
| 21:41.40 | *** join/#asterisk-dev leedm777 (~textual@50-81-28-152.client.mchsi.com) |
| 21:41.40 | *** join/#asterisk-dev Tim_Toady (~fuzzy@snf-33276.vm.okeanos.grnet.gr) |
| 21:42.50 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-82-81-175-197.red.bezeqint.net) |
| 21:42.50 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-fqjzoroktzqzoauf) |
| 21:42.50 | *** join/#asterisk-dev newtonr (RustyNewto@nat/digium/x-gbteknbionqymgdl) |
| 21:42.50 | *** join/#asterisk-dev opticron (~opticron@75.76.59.35) |
| 21:42.50 | *** join/#asterisk-dev seanbright (~sean@asterisk/contributor-and-bug-marshal/seanbright) |
| 21:42.50 | *** mode/#asterisk-dev [+oooo tzafrir newtonr opticron seanbright] by asimov.freenode.net |
| 21:57.33 | *** join/#asterisk-dev X-Rob (X-Rob@gateway/web/irccloud.com/x-uzzxcmcfrbbqejgm) |
| 22:05.51 | *** join/#asterisk-dev GameGamer43 (sid5533@gateway/web/irccloud.com/x-jumibvvdlikcyjpx) |
| 22:18.37 | *** join/#asterisk-dev coreyfarrell (~coreyfarr@24-177-86-98.dhcp.nwtn.ct.charter.com) |
| 22:50.02 | otherwiseguy | file: getting ready to push a few fixes to yappcap. Looks like it might still be used in testsuite? Don't think I broke anything, but just figured I'd let you know just in case. |
| 22:50.18 | otherwiseguy | also: HI |
| 22:50.27 | file | otherwiseguy, HI! YES! HI! |
| 22:51.18 | otherwiseguy | file: apparently more recent Cython's end up incrementing the refcount of something that could be PyObject * which could be NULL. Segfaults are bad. |
| 22:51.34 | file | nah, they're great |
| 22:52.27 | otherwiseguy | Also, spelling errors and using cythonize() to generate the yappcap.c file to make building a tiny bit less ugly. |
| 22:52.50 | otherwiseguy | Still including a yappcap.c file in the checkout, so doubt that'll make any difference. |
| 22:52.59 | file | practice reckless optimism |
| 22:53.09 | otherwiseguy | But, you never know. So I thought I'd bring it up. :p |
| 22:53.26 | otherwiseguy | I can make a tag for you guys to check out pre-change if you want. :) |
| 22:53.42 | file | I'm not that bothered |
| 22:54.01 | otherwiseguy | There's a small chance we may actually use it for some tests in openstack, so I figured I'd give it a little bit of attention. :) |
| 22:54.16 | otherwiseguy | I should probably add some tests to yappcap. ;) |
| 23:26.46 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-fqjzoroktzqzoauf) |
| 23:31.45 | *** join/#asterisk-dev leedm777 (~textual@216.186.181.138) |
| 23:46.37 | *** join/#asterisk-dev klow (~textual@c-98-247-49-57.hsd1.wa.comcast.net) |