| 08:09.30 | *** join/#asterisk-dev infobot (ibot@rikers.org) |
| 08:09.30 | *** topic/#asterisk-dev is Asterisk Development Discussion -=- http://www.asterisk.org/developers -=- Tier 2 and 3.14159265 support is in #asterisk -=- Check out our blog! asterisk-blog.com -=- Follow on Twitter at @AsteriskDev |
| 08:43.27 | *** join/#asterisk-dev luckman212 (~luckman21@unaffiliated/luckman212) |
| 08:44.45 | *** join/#asterisk-dev anthm (~anthm@freeswitch/developer/anthm) |
| 08:49.29 | *** join/#asterisk-dev GameGamer43 (sid5533@gateway/web/irccloud.com/x-ypyfdzrasnmbwrlz) |
| 12:21.53 | *** join/#asterisk-dev acidfu_ (~acidfoo@198-48-218-90.cpe.pppoe.ca) |
| 12:31.02 | *** join/#asterisk-dev newtonr (~newtonr@173-21-146-94.client.mchsi.com) |
| 12:31.02 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 12:47.43 | *** join/#asterisk-dev WIMPy (~wimpy@85.183.95.26) |
| 13:09.20 | *** join/#asterisk-dev putnopvut (~putnopvut@asterisk/master-of-queues/mmichelson) |
| 13:09.21 | *** mode/#asterisk-dev [+o putnopvut] by ChanServ |
| 14:02.25 | *** join/#asterisk-dev cresl1n (Adium@asterisk/libpri-and-libss7-expert/Cresl1n) |
| 14:02.25 | *** mode/#asterisk-dev [+o cresl1n] by ChanServ |
| 14:15.20 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-omhuzjvusuvkqpeo) |
| 14:26.40 | cresl1n | waves |
| 14:52.29 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-renngkkspsrtjhzq) |
| 15:10.37 | coreyfarrell | file: well the job succeeded, logs.tar.gz is 1.11gb. how close is that to the limit? |
| 15:24.37 | gtjoseph | anyone tried packaging asterisk as a snap yet? |
| 15:37.40 | cresl1n | snap? |
| 15:37.42 | cresl1n | what's that |
| 15:38.16 | cresl1n | oh that's the new ubuntu thing |
| 15:48.34 | gtjoseph | putnopvut: camels slaughtered. |
| 15:48.44 | rmudgett | gtjoseph: goats too? |
| 15:48.54 | gtjoseph | no, i like goats |
| 15:49.15 | putnopvut | excellent |
| 15:49.20 | putnopvut | I'll re-review |
| 16:04.48 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-179-40-172.cust.bezeqint.net) |
| 16:04.48 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 16:25.24 | *** join/#asterisk-dev F2Knight (~F2Knight@c-50-139-85-237.hsd1.or.comcast.net) |
| 17:09.02 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.cust.se.alltele.net) |
| 17:09.02 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 17:53.44 | putnopvut | There've been some periodic testsuite failures where we've seen MALLOC_DEBUG report that there is detected memory corruption of a freed JSON object. |
| 17:54.08 | *** join/#asterisk-dev pchero (~pchero@109.70.54.56) |
| 17:54.45 | putnopvut | I'm not able to reproduce this on my box, yet. |
| 17:55.06 | putnopvut | So I figure my best bet is to make it so we can potentially get more information from the testsuite when the test runs. |
| 17:55.24 | putnopvut | My thought is to enable the "memory backtrace on" CLI command on the test that has been observed to fail. |
| 17:55.55 | putnopvut | That way, the error from Asterisk will include the allocation backtrace, which could be very useful here. |
| 17:56.00 | *** join/#asterisk-dev klow (~textual@c-98-247-49-57.hsd1.wa.comcast.net) |
| 17:56.10 | putnopvut | Does anyone have any other good ideas for stuff I can add to help narrow down the problem space? |
| 18:02.11 | *** join/#asterisk-dev newtonr (~newtonr@173-21-146-94.client.mchsi.com) |
| 18:02.11 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 18:14.19 | file | coreyfarrell, it should be fine and since then I think I've added extra cleanup stuff to Jenkins |
| 18:43.42 | acidfu_ | is there any wiki/discussion started for AstriDevCon 2016 ? |
| 18:44.21 | acidfu_ | JunK-Y do you plan to go to Astricon this year ? :) |
| 18:45.40 | file | nothing as of yet |
| 18:49.04 | acidfu_ | ok ;) |
| 18:53.57 | JunK-Y | acidfu_: of course! You? |
| 18:54.03 | acidfu_ | yes :) |
| 18:59.26 | newtonr | file, I should probably update the wikipage for AstriDevCon 2016 |
| 18:59.36 | file | it does approach fastly |
| 19:00.20 | file | which is also the name of a CDN |
| 19:00.34 | *** join/#asterisk-dev tuxd00d (~tuxd00d@ip72-201-239-241.ph.ph.cox.net) |
| 19:01.32 | *** join/#asterisk-dev F2Knight (~F2Knight@c-50-139-85-237.hsd1.or.comcast.net) |
| 19:15.48 | *** join/#asterisk-dev klow (~textual@c-98-247-49-57.hsd1.wa.comcast.net) |
| 19:28.06 | gtjoseph | can someone pull up res_pjsip_pubsub for me and look at subscription_tree_destructor??? |
| 19:28.38 | gtjoseph | putnopvut: or rmudgett ^^ |
| 19:28.56 | putnopvut | Sure thing! |
| 19:29.37 | gtjoseph | i'm now seeing a segv in ao2_cleanup and i think it's the sub_tree destructor |
| 19:31.22 | putnopvut | Okay, what did you want me to look at in there? |
| 19:32.04 | gtjoseph | ast_sip_push_task_synchronous only checks that the thread is a server thread, not that we're already on the same serializer... |
| 19:32.19 | gtjoseph | we also pass in sub_tree to the task |
| 19:33.18 | gtjoseph | this SHOULDN'T be an issue but I am getting a segv when ao2_cleanup runs on sub_tree |
| 19:33.37 | gtjoseph | an dunfortunately the bt stops at ao2A_cleanup |
| 19:34.19 | gtjoseph | anyway, looking for opinions as to whether we should check manually that we're already on the sub_tree's serializer |
| 19:34.53 | putnopvut | Hm |
| 19:35.25 | putnopvut | I wonder what would happen if that weren't synchronous |
| 19:35.42 | gtjoseph | ao2 would barf roaylly |
| 19:35.46 | putnopvut | heh |
| 19:36.20 | *** join/#asterisk-dev mjordan (mjordan@nat/digium/x-bvevjijkwsdzdhci) |
| 19:36.20 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 19:36.54 | gtjoseph | damn, just got the segv again on ao2_cleanup with no extra stack frames, even with dont_optimize and better_backtraces |
| 19:37.06 | putnopvut | geez |
| 19:38.01 | putnopvut | So, just going through things logically |
| 19:38.10 | gtjoseph | checks dmesg for system errors |
| 19:38.18 | putnopvut | The reason for pushing a task synchronously is that you want the task to complete before moving on |
| 19:38.36 | gtjoseph | just like it hsould |
| 19:38.38 | putnopvut | And we have that check for whether we're already in a servant thread in order to prevent deadlocks |
| 19:39.00 | putnopvut | Because if we are already in the serializer and push it into the same serializer...oops |
| 19:39.02 | putnopvut | But |
| 19:39.04 | gtjoseph | well, it's also a short-cut to just run the callback in-line |
| 19:39.15 | putnopvut | If we're in some other servant thread and want to push to a serializer...that might work |
| 19:39.34 | putnopvut | But I'm hesitant |
| 19:40.47 | gtjoseph | well, i'm going to compare ast_threadpool_serializer_get_current() to sub_tree->serializer at least and run the task in-line if we are. |
| 19:41.03 | putnopvut | All righty |
| 19:41.37 | gtjoseph | AH!!! When it does SEGV is in a monitor thread, not the serializer |
| 19:42.26 | gtjoseph | it's being called from pubsub_on_server_timeout IF the push_task fails. |
| 19:44.16 | gtjoseph | damn i wonder if sub_tree->serializer isn't valid when we try to push |
| 19:53.49 | rmudgett | gtjoseph: If you have MALLOC_DEBUG enabled you'd crash with 0xdeaddead as the address if sub_tree->serializer is not valid. |
| 19:54.12 | gtjoseph | yeah good point. testing now. |
| 19:54.52 | gtjoseph | can't wait to see what else it discovers. :) |
| 19:54.54 | rmudgett | Also you may have some memory corruption because of the backtraces being so less than informative |
| 19:59.38 | gtjoseph | no help. time to reboot |
| 20:02.39 | *** join/#asterisk-dev gtjoseph (~gtjoseph@unaffiliated/gtj) |
| 20:20.54 | *** join/#asterisk-dev raspberrypifan (~raspberry@cpe-69-204-233-226.nyc.res.rr.com) |
| 20:28.34 | gtjoseph | a client termination, a subscription timer and a transaction timer walk into a bar... |
| 20:28.43 | gtjoseph | finish the joke. |
| 20:29.22 | file | and then a patch goes up to fix the problem? |
| 20:29.36 | gtjoseph | no it don't |
| 20:39.17 | gtjoseph | rmudgett: you missed something in my review. how is that possible??? |
| 20:39.56 | rmudgett | What'd I miss? |
| 20:40.39 | gtjoseph | there was at least 1 place where i was declaring sub_tree without initializing it, but testing it for NULL. :) |
| 20:40.50 | gtjoseph | that was the cause of my SEGVs this afternoon. |
| 20:41.43 | gtjoseph | pubsub_on_server_timeout. and the order of the predicates was wrong. |
| 20:42.57 | rmudgett | I still gave it a -1. :D |
| 20:43.11 | gtjoseph | yeah ok |
| 20:52.26 | rmudgett | Though why you like pulling assignments into if tests to make complicated expressions... |
| 20:53.46 | gtjoseph | just to give you a headache |
| 20:54.15 | gtjoseph | and because if i separated them someone would complain that i could have combined them |
| 20:55.14 | gtjoseph | tell you what. i'll separate them before posting again. |
| 20:59.21 | gtjoseph | rmudgett: i just remembered why sub_tree->pending_terminates needs to above 1... look at pubsub_on_evsub_state again... |
| 21:01.43 | rmudgett | What about it. It tests if it is zero before incrementing it. There aren't any places where it is incremented without prevously testing if it is zero. |
| 21:02.04 | rmudgett | It thus should not be able to get incremented above 1 |
| 21:02.30 | gtjoseph | yes. and that's the wrong part. it needs to be above 1. |
| 21:02.38 | gtjoseph | i need to increment before the test |
| 21:02.53 | gtjoseph | and only push if it's == 1 |
| 21:03.54 | gtjoseph | crap. i need to break that function into 2 |
| 21:42.54 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.cust.se.alltele.net) |
| 21:42.54 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 21:43.59 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.cust.se.alltele.net) |
| 21:43.59 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 22:00.32 | *** join/#asterisk-dev klow (~textual@c-98-247-49-57.hsd1.wa.comcast.net) |
| 22:04.11 | *** join/#asterisk-dev klow (~textual@c-98-247-49-57.hsd1.wa.comcast.net) |
| 22:08.31 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.cust.se.alltele.net) |
| 22:08.31 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 22:12.08 | *** join/#asterisk-dev shido6 (~shido6@CPE00fc8d94d3f3-CM00fc8d94d3f0.cpe.net.cable.rogers.com) |
| 22:22.24 | *** join/#asterisk-dev raspberrypifan (~raspberry@2604:2000:6016:be00:6355:b61:14e7:dbba) |
| 23:07.55 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-omhuzjvusuvkqpeo) |
| 23:09.28 | *** join/#asterisk-dev tuxd00d (~tuxd00d@ip72-201-239-241.ph.ph.cox.net) |
| 23:11.59 | *** join/#asterisk-dev klow (~textual@c-98-247-49-57.hsd1.wa.comcast.net) |
| 23:41.57 | *** join/#asterisk-dev tuxd00d (~tuxd00d@ip72-201-239-241.ph.ph.cox.net) |