| 00:20.22 | *** join/#asterisk-dev pabelanger (~pabelange@asterisk/contributor-and-bug-marshal/pabelanger) |
| 00:20.22 | *** mode/#asterisk-dev [+o pabelanger] by ChanServ |
| 00:53.20 | *** join/#asterisk-dev GameGamer43 (uid5533@gateway/web/irccloud.com/x-dyjacoeluhcewtej) |
| 01:20.20 | *** join/#asterisk-dev serafie (~erin@76.73.167.231) |
| 01:38.53 | *** join/#asterisk-dev file (~file@asterisk/developer-and-muffin-lover/file) |
| 01:38.53 | *** mode/#asterisk-dev [+o file] by ChanServ |
| 01:57.58 | *** join/#asterisk-dev opticron (~opticron@pianoben.ch) |
| 01:57.58 | *** mode/#asterisk-dev [+o opticron] by ChanServ |
| 02:09.35 | *** join/#asterisk-dev mjordan (~mjordan@216.186.152.216) |
| 02:09.35 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 02:33.19 | *** join/#asterisk-dev pabelanger (~pabelange@asterisk/contributor-and-bug-marshal/pabelanger) |
| 02:33.20 | *** mode/#asterisk-dev [+o pabelanger] by ChanServ |
| 03:33.12 | *** join/#asterisk-dev sruffell (~sruffell@asterisk/the-kernel-guy/sruffell) |
| 03:33.12 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 05:33.04 | *** join/#asterisk-dev GameGamer43 (uid5533@gateway/web/irccloud.com/x-gtvazyvkwrroesgs) |
| 06:54.25 | *** join/#asterisk-dev jmls (~somefake@77.107.171.82) |
| 07:22.31 | snuff-work | posted patch wedhorn.. https://issues.asterisk.org/jira/browse/ASTERISK-21066 |
| 08:09.14 | wedhorn | yep, saw it |
| 08:37.41 | *** join/#asterisk-dev hehol (~hehol@2001:1438:1009:200:34b9:b571:25dd:f1c) |
| 08:40.59 | *** join/#asterisk-dev bulkorok (~bulkorok@85.183.36.36) |
| 09:06.05 | *** join/#asterisk-dev tm1000 (tm1000@2600:3c01::f03c:91ff:fe70:fce6) |
| 09:47.44 | *** join/#asterisk-dev schmidts (~schmidts@vie-188-118-255-190.dsl.sil.at) |
| 09:47.46 | schmidts | good morning |
| 11:36.43 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 11:36.44 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 11:51.51 | *** join/#asterisk-dev oej (~Adium@2001:16d8:cc57:1000::42:1005) |
| 11:51.51 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 12:00.08 | *** join/#asterisk-dev sgimeno (~sgimeno@163.117.206.10) |
| 12:09.33 | *** join/#asterisk-dev sruffell (~sruffell@asterisk/the-kernel-guy/sruffell) |
| 12:09.33 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 12:28.48 | *** join/#asterisk-dev saghul (~saghul@ip3e830637.speed.planet.nl) |
| 12:36.05 | tzafrir | Author: kharwell not defined in AUTHORS file |
| 12:54.17 | mjordan | tzafrir: repotools should be updated |
| 12:57.59 | *** join/#asterisk-dev barbosa2 (~juliano.b@177.157.83.193) |
| 12:59.41 | *** join/#asterisk-dev barbosa2 (~juliano.b@177.157.83.193) |
| 13:04.44 | *** join/#asterisk-dev serafie (~erin@76.73.167.231) |
| 13:08.17 | oej | Afternoon! |
| 13:12.13 | Marquis42 | oej: How is the future? ;) |
| 13:12.46 | oej | Well, here in the future we have no cars. ASterisk-powered small jet-powered flights |
| 13:12.56 | oej | Asterisk-powered tea vending machines |
| 13:13.05 | oej | And rack servers that fail - still. |
| 13:13.22 | Marquis42 | heh... The more things change, the more they stay the same. :) |
| 13:43.26 | *** join/#asterisk-dev barbosa2 (~juliano.b@177.157.83.193) |
| 13:55.40 | *** join/#asterisk-dev serafie (~erin@nat/digium/x-jnakkayubusyvauu) |
| 13:56.41 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 13:56.46 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 14:26.38 | *** join/#asterisk-dev Juggie (~Juggie@unaffiliated/juggie) |
| 14:33.17 | *** join/#asterisk-dev mjordan (~mjordan@nat/digium/x-pacsnnbhjzerfyrx) |
| 14:33.17 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 14:37.20 | *** join/#asterisk-dev leedm777 (~leedm777@nat/digium/x-uuwrpdezejacterl) |
| 15:01.48 | *** join/#asterisk-dev pabelanger (~pabelange@asterisk/contributor-and-bug-marshal/pabelanger) |
| 15:01.48 | *** mode/#asterisk-dev [+o pabelanger] by ChanServ |
| 15:16.52 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 15:16.52 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 15:32.32 | *** join/#asterisk-dev putnopvut (~putnopvut@asterisk/master-of-queues/mmichelson) |
| 15:32.32 | *** mode/#asterisk-dev [+o putnopvut] by ChanServ |
| 16:14.43 | acidfoo | 1.8.20.1 is freezing here, and I can reproduce it, I have a core show locks result |
| 16:17.20 | seanbright | edit asterisk.conf |
| 16:17.22 | seanbright | put: |
| 16:17.26 | seanbright | freezing=no |
| 16:17.30 | seanbright | and then restart |
| 16:18.49 | pabelanger | .notbad |
| 16:19.00 | acidfoo | seanbright, when I do that, asterisk just doesnt start :P |
| 16:19.11 | seanbright | hmmm, add: |
| 16:19.13 | seanbright | startup=yes |
| 16:19.32 | acidfoo | it says that there are incompatible ;P |
| 16:19.37 | acidfoo | they are* |
| 16:19.58 | acidfoo | anyway I look into the trace and report it |
| 16:21.40 | acidfoo | I'll* |
| 16:21.43 | acidfoo | dang, |
| 16:30.24 | acidfoo | http://postimage.org/image/d1ekpsrwb/ |
| 16:30.29 | acidfoo | http://pastebin.com/7HUifMVK |
| 16:30.49 | acidfoo | trace for a generated backtrace |
| 16:30.59 | acidfoo | gdb -ex "thread apply all bt" --batch /usr/sbin/asterisk $(pidof asterisk) > bt-threads.txt |
| 16:31.02 | acidfoo | with that |
| 16:31.18 | acidfoo | I just pasted the to process that were locked together |
| 16:33.01 | acidfoo | http://pastebin.com/q6pSA4Ed |
| 16:33.11 | acidfoo | the two locked process from core show locks |
| 16:36.07 | *** join/#asterisk-dev greenwolf (42570420@gateway/web/freenode/ip.66.87.4.32) |
| 16:38.50 | *** join/#asterisk-dev hexanol (~bibi@modemcable094.94-70-69.static.videotron.ca) |
| 16:38.54 | acidfoo | yo hexanol |
| 16:39.07 | hexanol | who are you ? |
| 16:39.11 | hexanol | ;) |
| 16:39.14 | acidfoo | bleh :P |
| 16:51.01 | *** join/#asterisk-dev oej (~Adium@2001:16d8:cc57:1000::42:1003) |
| 16:51.01 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 17:15.13 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 17:15.13 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 17:18.10 | acidfoo | static int show_channels_cb(void *__cur, void *__arg, int flags) |
| 17:18.13 | acidfoo | in chan_sip |
| 17:18.21 | acidfoo | some lock have been removed here |
| 17:22.15 | acidfoo | food time |
| 17:43.45 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 17:43.45 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 17:54.14 | acidfoo | erm, not removed ,but added |
| 18:01.11 | acidfoo | i'll test if that problem was present in 1.8.17 |
| 18:11.48 | *** join/#asterisk-dev anthm (~anthm@freeswitch/developer/anthm) |
| 18:14.40 | *** join/#asterisk-dev w9sh (~w9sh@50-79-224-193-static.hfc.comcastbusiness.net) |
| 18:26.20 | *** join/#asterisk-dev moy (~moy@173.239.155.74) |
| 18:43.35 | acidfoo | same problem with 1.8.18.0 |
| 18:49.48 | mjordan | I vaguely remember that we have an issue open for a deadlock occurring when you do a 'sip show channels' |
| 18:50.44 | mjordan | yup! |
| 18:50.45 | mjordan | https://issues.asterisk.org/jira/browse/ASTERISK-20550 |
| 18:52.25 | acidfoo | i'm not able to freeze 1.8.17. but I can freeze 1.8.18 and 1.8.20.1 |
| 18:53.11 | acidfoo | everytime |
| 18:53.37 | acidfoo | a python script sends tons of 'ami command sip show channels' |
| 18:53.55 | mjordan | yup |
| 18:53.56 | acidfoo | and sipp send lots of register (and there is a voicemail) |
| 18:53.59 | mjordan | yup |
| 18:54.01 | acidfoo | so it generates mwi stuff |
| 18:54.02 | mjordan | :-) |
| 18:54.11 | acidfoo | good ;) |
| 18:54.45 | mjordan | the locking inversion happens in parse_register_contact. We don't have a patch yet |
| 18:54.58 | mjordan | but it will only occur when you use that CLI command |
| 18:55.35 | mjordan | well... probably. There may be other places, but the CLI command is much more likely to do it as it iterates over all dialogs in the container |
| 18:55.43 | mjordan | locking the container then the dialogs. |
| 18:56.06 | mjordan | whereas, the register code is doing lock pvt => lock container => |
| 18:56.07 | acidfoo | I feel like this commit didn't help http://lists.digium.com/pipermail/asterisk-commits/2012-September/056919.html |
| 18:56.16 | acidfoo | yeah |
| 18:56.37 | mjordan | maybe, but I know why he committed that, and it was very much needed |
| 18:56.55 | acidfoo | yeah I don't know, I'm just trying to help =) |
| 18:57.05 | mjordan | there was an incredibly evil bug where a thread would get switched out of context in the middle of parsing a request, and when it came back the data was no longer what it thought it was |
| 18:57.40 | mjordan | well, close to that at any rate. The parsing of the request would not be finished and other things would attempt to access it. |
| 18:57.57 | mjordan | acidfoo: why are you using the CLI command so often? |
| 18:58.23 | mjordan | not that you shouldn't be able to, just curious |
| 18:58.23 | acidfoo | to freeze asterisk ? ^ ^ |
| 18:58.26 | acidfoo | ;P |
| 18:58.30 | mjordan | heh |
| 18:58.38 | acidfoo | nah, monin is doing it |
| 18:58.40 | acidfoo | munin |
| 18:58.47 | acidfoo | plugin to graph |
| 18:58.51 | mjordan | ah. |
| 18:58.58 | mjordan | :-( |
| 18:59.08 | acidfoo | at the bottom of the commit I showed you |
| 18:59.14 | mjordan | CLI command over AMI always make me sad. |
| 18:59.23 | acidfoo | around that part: @@ -18485,8 +18507,11 @@ |
| 18:59.27 | acidfoo | does it was to fix anything ? |
| 18:59.38 | acidfoo | the stuff added in the show_channels_cb() function |
| 19:00.27 | acidfoo | but you're right about using command over AMI |
| 19:00.39 | mjordan | well, you probably should do that, because you don't own the pvt |
| 19:01.00 | acidfoo | yes but someone else is iterating around dialogs at the same time |
| 19:01.06 | acidfoo | and lock each pvt |
| 19:01.22 | mjordan | yes, but the locking order should be container => pvt. |
| 19:01.28 | mjordan | it's probably the register code that should change. |
| 19:01.34 | acidfoo | http://pastebin.com/7HUifMVK |
| 19:01.39 | mjordan | if it needs the pvt, it should unlock it prior to locking the container. |
| 19:01.42 | acidfoo | http://postimage.org/image/d1ekpsrwb/ |
| 19:01.44 | mjordan | ref bump it first |
| 19:02.14 | mjordan | right, but the show channels callback is actually doing it correct by locking the container, then the pvt |
| 19:02.19 | acidfoo | yes |
| 19:02.23 | acidfoo | ok |
| 19:02.25 | acidfoo | erm |
| 19:02.37 | mjordan | it is the registration code that is borked, because it's making a new pvt while it holds the lock to a current one, and making a new pvt entails locking the container |
| 19:02.38 | acidfoo | it doesn't lock the dialogs first |
| 19:02.44 | acidfoo | ok |
| 19:02.51 | acidfoo | you know better :) |
| 19:03.03 | mjordan | heh |
| 19:03.06 | acidfoo | and I'm happy you're giving me feedback ;) |
| 19:03.12 | mjordan | when chan_sip is involved, it's all touchy |
| 19:03.24 | acidfoo | ok heh:) |
| 19:03.38 | acidfoo | but yeah, those new locks in show_channels_cb() created a crash |
| 19:03.41 | acidfoo | a freeze |
| 19:03.43 | mjordan | yup |
| 19:03.49 | acidfoo | so is it a regression ? :P |
| 19:03.51 | mjordan | sure |
| 19:03.54 | mjordan | :-) |
| 19:04.21 | mjordan | Feel free to mark it as such on the issue |
| 19:04.57 | mjordan | I can see if that bumps it higher in the queue. Honestly however, it typically is going to impact systems that are doing behavior that is really, well, wrong. |
| 19:05.04 | mjordan | (not that we shouldn't fix it however0 |
| 19:05.05 | mjordan | ) |
| 19:05.17 | acidfoo | and do you think that those lock added in show_channels_cb() was fixing something ? because meanwhile I'll run 1.8.20 without them |
| 19:05.53 | mjordan | they prevent a situation where you would attempt to show a pvt when it was being altered, getting either incorrect data or worse |
| 19:06.04 | mjordan | but that's most likely a rather rare event. |
| 19:06.19 | acidfoo | probably rarer than the situation it introduce ;/ |
| 19:06.22 | mjordan | probably |
| 19:06.32 | acidfoo | ok good |
| 19:07.54 | acidfoo | thank you for your time |
| 19:08.38 | mjordan | np - hanging out in here is what we do :-0 |
| 19:08.40 | mjordan | :-) |
| 19:43.47 | *** join/#asterisk-dev jmls (~somefake@77.107.171.82) |
| 19:46.09 | *** part/#asterisk-dev hexanol (~bibi@modemcable094.94-70-69.static.videotron.ca) |
| 21:57.12 | elguero | :q |
| 22:05.42 | leifmadsen | elguero: I approve of your choice of editor |
| 22:06.30 | mjordan | eclipse? |
| 22:06.31 | mjordan | hides |
| 22:08.48 | file | *hiss* |
| 22:10.08 | elguero | hehe |
| 22:13.54 | mjordan | YAML markup ftw |
| 22:14.53 | mjordan | whenever I run svn blame on chan_sip, I always pipe it to a file named 'blame_chan_sip.txt'. It makes me feel good. |
| 22:17.27 | file | mjordan, run cppcheck on chan_sip.c |
| 22:17.38 | file | :D |
| 22:21.29 | mjordan | does cppcheck grab a bottle of scotch, hide under its desk, and cry itself to sleep? |
| 22:21.56 | file | YES! wait, no. |
| 22:55.12 | seanbright | sruffell: grrrrrrrrrrr |
| 22:55.14 | seanbright | wctc4xxp 0000:06:08.0: Failed to create channel in timeslot 36. Response from DTE was (ffbd). |
| 22:55.18 | seanbright | ffbd! |
| 22:55.48 | seanbright | i need to make that stand for something... f*ckity f*ck bad dsp |
| 23:06.42 | sruffell | seanbright: I know datachomper was looking into that a couple of weeks ago. I think he has a patch on JIRA to mark that timeslot bad, but not to reload yet when all the channels are closed. |
| 23:06.51 | sruffell | but I hate to frown and run, but :( |
| 23:12.17 | *** join/#asterisk-dev Wildblue (~male@c-69-140-154-17.hsd1.md.comcast.net) |
| 23:12.29 | snuff-work | mm.. which person is doing the bridging api stuff again.. |
| 23:13.35 | file | Richard Mudgett, who does not frequent IRC and jrose_atDigium |
| 23:14.06 | snuff-work | i still haven't chased down my softmix issue yet.. :( |
| 23:17.11 | snuff-work | the work converting stuff looks really nice though props ppl :) |
| 23:18.04 | *** part/#asterisk-dev Wildblue (~male@c-69-140-154-17.hsd1.md.comcast.net) |
| 23:38.20 | *** part/#asterisk-dev mjordan (~mjordan@nat/digium/x-pacsnnbhjzerfyrx) |