| 02:10.42 | *** join/#asterisk-dev jameswf (uid27319@gateway/web/irccloud.com/x-vhpgydvexexfnifc) |
| 02:10.42 | *** join/#asterisk-dev TazzNZ (~TazzNZ@office-nat1.ytechops.co.nz) |
| 02:10.42 | *** join/#asterisk-dev pabelanger (~pabelange@asterisk/contributor-and-bug-marshal/pabelanger) |
| 02:10.42 | *** join/#asterisk-dev jrose_atDigium (~jrose@nat/digium/x-kvrwktqrlykxcmvd) |
| 02:10.42 | *** join/#asterisk-dev zerohalo (~zerohalo@2601:6:f80:224:acbf:d145:ab95:9423) |
| 02:10.42 | *** join/#asterisk-dev Stefan27 (~Stefan27@212.247.4.149) |
| 02:10.42 | *** join/#asterisk-dev Qwell (north@pdpc/sponsor/digium/Qwell) |
| 02:10.42 | *** join/#asterisk-dev Gugge (gugge@194.150.109.208) |
| 02:10.42 | *** join/#asterisk-dev felipe_ (~felipe@unaffiliated/felipe) |
| 02:10.42 | *** join/#asterisk-dev X-Rob (X-Rob@gateway/web/irccloud.com/x-boeazzlbzlokrgfu) |
| 02:10.42 | *** join/#asterisk-dev wedhorn (~wedhorn@110-175-181-222.static.tpgi.com.au) |
| 02:10.42 | *** join/#asterisk-dev Juggie (~Juggie@unaffiliated/juggie) |
| 02:10.42 | *** join/#asterisk-dev snuff-work (~snuffy@210.8.167.13) |
| 02:10.42 | *** join/#asterisk-dev Sjors (~sgielen@foo.kassala.de) |
| 02:10.42 | *** join/#asterisk-dev leedm777 (~leedm777@nat/digium/x-hncdnfiaearlyvfm) |
| 02:10.42 | *** join/#asterisk-dev angler (~angler@pdpc/sponsor/digium/angler) |
| 02:10.42 | *** join/#asterisk-dev seanbright (~sean@asterisk/contributor-and-bug-marshal/seanbright) |
| 02:10.42 | *** mode/#asterisk-dev [+oooo pabelanger Qwell snuff-work seanbright] by sinisalo.freenode.net |
| 02:10.42 | *** join/#asterisk-dev ivan` (~ivan@unaffiliated/ivan/x-000001) |
| 02:10.42 | *** join/#asterisk-dev wdoekes (~walter@wjd.osso.nl) |
| 02:10.42 | *** join/#asterisk-dev tm1000 (tm1000@2600:3c00::f03c:91ff:fe6e:6894) |
| 02:10.42 | *** join/#asterisk-dev dwayne (~dwayne@c-71-207-208-112.hsd1.al.comcast.net) |
| 02:10.43 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-218-28-58.cablep.bezeqint.net) |
| 02:10.43 | *** join/#asterisk-dev tonyclewis (sid6025@gateway/web/irccloud.com/x-ndewrkjihzdjhqim) |
| 02:10.43 | *** join/#asterisk-dev pc-m (~pcm@modemcable094.94-70-69.static.videotron.ca) |
| 02:10.43 | *** join/#asterisk-dev coreyfarrell (~cfarrell@2001:4830:117f:2:74f3:70c1:1710:dfc) |
| 02:10.43 | *** join/#asterisk-dev hexanol (~bibi@modemcable094.94-70-69.static.videotron.ca) |
| 02:10.43 | *** join/#asterisk-dev drmessano (~nonya@pdpc/supporter/active/drmessano) |
| 02:10.43 | *** join/#asterisk-dev gtjoseph (~gtj0@unaffiliated/gtj) |
| 02:10.43 | *** join/#asterisk-dev GameGamer43 (sid5533@gateway/web/irccloud.com/x-iaqkfvkcukahhmhd) |
| 02:10.43 | *** join/#asterisk-dev russellb (~russellb@redhat/russellb) |
| 02:10.43 | *** join/#asterisk-dev mbowie (~mbowie@162.212.36.9) |
| 02:10.43 | *** join/#asterisk-dev jsmith (~jsmith@fedora/jsmith) |
| 02:10.43 | *** join/#asterisk-dev HeN (uid3747@gateway/web/irccloud.com/x-swrxlvephqozwzvn) |
| 02:10.43 | *** join/#asterisk-dev WIMPy (~wimpy@e183095026.adsl.alicedsl.de) |
| 02:10.43 | *** join/#asterisk-dev opticron (~opticron@75.76.148.219) |
| 02:10.43 | *** mode/#asterisk-dev [+oooo tzafrir russellb jsmith opticron] by sinisalo.freenode.net |
| 02:10.43 | *** join/#asterisk-dev moy (~moy@198.199.91.192) |
| 02:10.43 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 02:10.43 | *** join/#asterisk-dev spditner (~simon@76-10-182-232.dsl.teksavvy.com) |
| 02:10.43 | *** join/#asterisk-dev skrusty (~skrusty@168.63.14.171) |
| 02:10.43 | *** join/#asterisk-dev chazzam (~chazz@donutokyo.info) |
| 02:10.43 | *** join/#asterisk-dev datachomper (~russ@nat/digium/x-xblmwvkpkhveodfd) |
| 02:10.43 | *** join/#asterisk-dev Tim_Toady (~fuzzy@83.212.108.130) |
| 02:10.43 | *** join/#asterisk-dev file (~file@asterisk/developer-and-muffin-lover/file) |
| 02:10.43 | *** mode/#asterisk-dev [+oo malcolmd file] by sinisalo.freenode.net |
| 02:20.14 | *** join/#asterisk-dev ChanServ (ChanServ@services.) |
| 02:20.14 | *** mode/#asterisk-dev [+o ChanServ] by sinisalo.freenode.net |
| 02:22.01 | *** join/#asterisk-dev luckman212 (~luckman21@unaffiliated/luckman212) |
| 02:25.21 | *** join/#asterisk-dev elguero (~miguel323@2001:470:1f06:12c4::2) |
| 02:41.32 | *** join/#asterisk-dev coreyfarrell (~cfarrell@2001:4830:117f:2:d15d:4aa2:a953:122f) |
| 03:12.33 | *** join/#asterisk-dev SGjunior (~sgjunior@out-pq-147.wireless.telus.com) |
| 03:20.46 | *** join/#asterisk-dev anhvt (~anhvt@113.190.40.214) |
| 03:23.26 | *** join/#asterisk-dev sruffell (~sruffell@asterisk/the-kernel-guy/sruffell) |
| 03:23.26 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 03:27.00 | *** join/#asterisk-dev GameGamer43 (sid5533@gateway/web/irccloud.com/x-ultelrjbtsovsmqc) |
| 05:16.21 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@64.235.216.2) |
| 06:33.42 | *** join/#asterisk-dev bulkorok (~Benjamin@85.183.61.47) |
| 06:55.01 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 06:55.02 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 07:09.18 | *** join/#asterisk-dev otherwiseguy (~otherwise@63-254-35-62.ip.mcleodusa.net) |
| 07:39.24 | *** join/#asterisk-dev hehol (~hehol@2001:1438:1009:200:e13d:55be:76ee:ca83) |
| 09:44.23 | *** join/#asterisk-dev HeN (uid3747@gateway/web/irccloud.com/x-pmvnauczqyaanlcv) |
| 11:40.53 | *** join/#asterisk-dev sgriepentrog (~sgriepent@nat/digium/x-plxqevqntkrglnoq) |
| 12:04.09 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@216-191-106-163.dedicated.allstream.net) |
| 12:18.55 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 12:19.11 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 12:52.25 | *** join/#asterisk-dev lirakis (~lirakis@50-198-204-238-static.hfc.comcastbusiness.net) |
| 13:02.28 | file | Friday! |
| 13:05.51 | lirakis | hey file |
| 13:05.55 | lirakis | so im looking at that issue |
| 13:06.14 | lirakis | and i think asterisk may have changed their SDP between .9 and .11 |
| 13:06.17 | lirakis | i think they did it correctly |
| 13:06.23 | file | yes |
| 13:06.28 | lirakis | but there is a thread on the webrtc google group related to it |
| 13:06.48 | lirakis | kk |
| 13:06.51 | lirakis | just verifying |
| 13:07.01 | lirakis | it is actually broken with both chrome and firefox ATM |
| 13:07.08 | lirakis | they are broken that is |
| 13:07.12 | file | who? |
| 13:07.16 | lirakis | i guess .11 is still rc1 |
| 13:07.34 | file | .11 was released and when I committed the WebRTC changes last week it was working for both Firefox+Chrome inbound and outbound |
| 13:07.41 | lirakis | neither chrome nor FF parse the m audio line from asterisk .11 |
| 13:07.43 | file | and others have said it's working for them |
| 13:08.02 | lirakis | if they use sipml5 or some thing that hacks the bajeezus out of the sdp ... probably |
| 13:08.13 | lirakis | but the browsers dont parse the SDP properly |
| 13:08.18 | lirakis | https://groups.google.com/forum/#!topic/sip_js/sqexozWaCL8 |
| 13:08.27 | lirakis | thats the reported issue - tested .11 against chrome and FF |
| 13:08.39 | lirakis | i believe its related to https://code.google.com/p/webrtc/issues/detail?id=2796 |
| 13:08.42 | lirakis | on the webrtc group |
| 13:09.02 | lirakis | so we need to basically work some temporary solution till the browsers roll out support |
| 13:09.42 | file | if you set force_avp it'll make the Asterisk SDP use RTP/SAVPF |
| 13:09.48 | file | provided avpf is also enabled |
| 13:10.29 | lirakis | ok - we can add that to our guide as well - but we will look into provider a solution within sip.js (we have a mediaHacks object) |
| 13:10.52 | lirakis | file, thanks for the clarification - sorry i didnt have the info yesterday |
| 13:11.11 | file | all the knobs exist now :D and hopefully if things change they can just be turned off... and then the world carries on |
| 13:11.27 | lirakis | heh yep |
| 13:11.29 | lirakis | thanks |
| 13:11.32 | file | cheers |
| 13:22.04 | *** join/#asterisk-dev hehol (~hehol@217.9.101.222) |
| 13:34.12 | opticron | another crash bug, hooray! |
| 14:02.38 | opticron | something in the bridging layer appears to be copying a frame without bumping the format refcount |
| 14:03.02 | file | is it not duping? |
| 14:03.29 | opticron | don't quite know yet |
| 14:03.37 | opticron | it's not bumping the ref, whatever it's doing |
| 14:04.27 | *** join/#asterisk-dev kharwell (~kharwell@nat/digium/x-uzmfwcsmdwimmiqy) |
| 14:12.42 | *** join/#asterisk-dev danjenkins_ (~dan@cpc11-folk2-2-0-cust171.1-2.cable.virginm.net) |
| 14:12.45 | *** part/#asterisk-dev lirakis (~lirakis@50-198-204-238-static.hfc.comcastbusiness.net) |
| 14:17.28 | *** join/#asterisk-dev Marquis42 (~mrbmw@d149-67-95-208.try.wideopenwest.com) |
| 14:19.47 | Marquis42 | Hrm, so newb question about the realtime API. Does it determine type somewhere down the line for columns that are INT vs. VARCHAR or whatever? I have an issue where 1.8 running against MySQL 5.6 can't update peers because it's trying to insert an empty string (i.e., '') instead of a NULL on an INT column. |
| 14:27.31 | *** join/#asterisk-dev newtonr (~newtonr@nat/digium/x-jjkgtckmwgkyjpbm) |
| 14:27.31 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 14:28.23 | seanbright | so what are the magic settings to make asterisk 11 work with chrome and webrtc? |
| 14:29.21 | seanbright | file: do you still have your test config from review #3679 laying around? |
| 14:30.06 | file | http://pastebin.com/bzUPbR9q |
| 14:30.19 | file | and I used the ast_tls_cert script in contrib/scripts to make the certs |
| 14:30.25 | seanbright | cool |
| 14:30.28 | seanbright | thanks boss |
| 14:32.45 | *** join/#asterisk-dev sruffell (~sruffell@asterisk/the-kernel-guy/sruffell) |
| 14:32.47 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 14:33.40 | seanbright | holy shit... it works |
| 14:33.44 | seanbright | file: you're the bestest. |
| 14:36.11 | file | seanbright, did you ever try out my audiohook patch? |
| 14:57.53 | *** join/#asterisk-dev jrose_atDigium (~jrose@nat/digium/x-ejvuuweptorhqtbt) |
| 14:59.58 | *** join/#asterisk-dev rmudgett (~rmudgett@nat/digium/x-tqtrdpsiwghcqazi) |
| 15:27.32 | seanbright | file: not yet |
| 15:27.42 | file | ack |
| 15:34.52 | *** join/#asterisk-dev otherwiseguy (~otherwise@nat/cisco/x-naylzgdjfkkuifpg) |
| 15:35.24 | *** join/#asterisk-dev danjenkins__ (~dan@nat/digium/x-qerluelvbfheockg) |
| 15:43.23 | *** join/#asterisk-dev newtonr (~newtonr@nat/digium/x-ibagghxlbrnqxosh) |
| 15:43.23 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 15:43.35 | hexanol | wondering, is it possible to use ast_manager_register_xml in a module that is compiled outside of asterisk ? and that the manager show commands shows correctly the documentation ? |
| 16:16.00 | *** join/#asterisk-dev mjordan (~mjordan@nat/digium/x-wcmjzhsngrvaydvc) |
| 16:16.00 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 16:22.22 | file | hexanol, not afaik |
| 16:41.08 | hexanol | ok |
| 16:54.14 | coreyfarrell | hexanol: there is a folder that you can drop the xml file with docs for your module. check out main/xmldoc.c - ast_xmldoc_load_documentation() |
| 16:54.49 | coreyfarrell | hexanol: I've never actually tried it, usually I just add any custom modules to the asterisk source tree and compile at the same time as asterisk |
| 17:12.24 | tm1000 | file: mjordan sooooo â12.4.0â is the magical version that solves world peace and makes coffee right? |
| 17:15.10 | file | tm1000, hi |
| 17:15.17 | tm1000 | file: hi |
| 17:15.58 | file | tm1000, how goes? |
| 17:16.07 | tm1000 | file: good how about you |
| 17:16.14 | file | tm1000, I'm super |
| 17:16.19 | tm1000 | file: have you solved world peace yet |
| 17:16.57 | rmudgett | The latest version always solves world peace. Its when people start using it that the peace is shattered. :) |
| 17:18.52 | mjordan | tm1000: I hope so |
| 17:23.10 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 17:23.10 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 17:25.00 | tm1000 | file: thanks for all your work on webrtc. i heard you love working on it |
| 17:25.03 | tm1000 | that and google voice |
| 17:25.16 | tm1000 | if we could just assign you to those two things the rest of your life iâd be happy |
| 17:25.29 | file | banishes tm1000 to a world where all he has is 12k MP3 files to listen to |
| 17:28.53 | opticron | wonders when the GV integration will finally die |
| 17:30.13 | jsmith | opticron: When everyone uses WebRTC support in Google Hangouts :-p |
| 17:30.35 | tm1000 | jsmith: probably |
| 17:30.48 | tm1000 | opticron: Im just waiting for the day where they turn it off |
| 17:30.58 | opticron | that's the date I'm suggesting |
| 17:36.12 | *** join/#asterisk-dev moy_ (~moy@198.199.91.192) |
| 17:43.12 | *** join/#asterisk-dev spditner (~simon@76-10-182-232.dsl.teksavvy.com) |
| 17:54.16 | *** join/#asterisk-dev opticron (~opticron@75.76.148.219) |
| 17:54.16 | *** mode/#asterisk-dev [+o opticron] by ChanServ |
| 17:56.22 | rmudgett | ast_translator_best_choice() leaks the joint_cap it allocates if there is no best codec |
| 18:00.43 | sruffell | Soâ¦I know there aren't any "official" packages for Ubuntu trusty listed at https://wiki.asterisk.org/wiki/display/AST/Asterisk+Packages, but is there a consensus about the best ones to use from the community are? |
| 18:00.53 | sruffell | oh, wait..I guess this is an astesrisk-users question. |
| 18:11.50 | *** join/#asterisk-dev WIMPy (~wimpy@e183095026.adsl.alicedsl.de) |
| 18:19.19 | *** join/#asterisk-dev ipengineer (~zconkle@static-71-252-134-63.dllstx.fios.verizon.net) |
| 18:36.46 | jameswf | some light reading... the case for using freeswitch.... http://pastebin.com/5J78YY4u and someone hug go hug TK |
| 18:38.00 | jameswf | oh and a little cherry on top... http://i.imgur.com/4s9XDWz.png |
| 18:43.10 | mbowie | lolz |
| 18:53.15 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@216-191-106-163.dedicated.allstream.net) |
| 18:58.01 | *** join/#asterisk-dev otherwiseguy (~otherwise@64.102.254.34) |
| 18:58.14 | *** join/#asterisk-dev oej (~Adium@87.96.134.129) |
| 18:58.14 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 19:09.39 | *** join/#asterisk-dev Jacoby6000 (~Jacoby600@static-71-252-134-63.dllstx.fios.verizon.net) |
| 19:14.31 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 19:14.32 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 19:15.33 | *** join/#asterisk-dev lirakis (~lirakis@50-198-204-238-static.hfc.comcastbusiness.net) |
| 19:16.04 | lirakis | hey file |
| 19:16.42 | *** join/#asterisk-dev anthm (~anthm@freeswitch/developer/anthm) |
| 19:16.44 | lirakis | we tested against asterisk .11 with force_avp and now asterisk is rejecting with 488 and the following |
| 19:16.47 | lirakis | [Jul 11 19:14:58] WARNING[7267][C-00000000]: chan_sip.c:10535 process_sdp: Rejecting secure audio stream without encryption details: audio 51250 RTP/SAVPF 111 103 104 0 8 106 105 13 126 |
| 19:18.56 | lirakis | this is against chrome stable |
| 19:20.17 | file | does your OpenSSL support DTLS-SRTP? |
| 19:21.38 | lirakis | hum good question - i didnt set up this test asterisk box, let me go check |
| 19:24.23 | lirakis | file, looks like it had 1.0.0 on it - updated and rebuilding for retest |
| 19:24.39 | file | I think it was added in... oh.. 1.0.1 something |
| 19:24.53 | lirakis | yea |
| 19:24.54 | file | <insert crazy OpenSSL versioning numbers here> |
| 19:24.59 | file | or letters. |
| 19:25.03 | lirakis | yeah he he |
| 19:25.11 | file | seanbright just got it working earlier ^_^ |
| 19:30.30 | *** part/#asterisk-dev hexanol (~bibi@modemcable094.94-70-69.static.videotron.ca) |
| 19:31.42 | lirakis | hum |
| 19:31.59 | lirakis | file, same error with OpenSSL 1.0.1e-fips 11 Feb 2013 |
| 19:32.08 | file | did you re-run configure? |
| 19:32.23 | mjordan | jameswf: good stuff. |
| 19:33.50 | lirakis | file, pretty sure the other dev did, but ill go run: make clean && ./configure && make && make install |
| 19:34.08 | file | the configure output will tell you if it detected srtp support in OpenSSL, and you can also check config.log - just search for srtp |
| 19:37.39 | opticron | so previously ast_format_rate returned 8000 on uninitialized format structs |
| 19:38.18 | opticron | it's new equivalent ast_format_get_sample_rate segfaults on null format pointers |
| 19:38.36 | mjordan | opticron: what is giving you a NULL format pointer? |
| 19:39.14 | opticron | the RTP layer |
| 19:39.17 | lirakis | im looking at the line where its complaining about - its looking for crypto lines for dtls? |
| 19:39.22 | lirakis | vs. fingerprint? |
| 19:39.25 | opticron | I think via a NULL RTP instance |
| 19:39.39 | mjordan | heh |
| 19:39.42 | mjordan | so many invalid things, so little time |
| 19:40.03 | mjordan | opticron: in the review I have up for unit tests, there is an ast_format_none that gets set on a channel |
| 19:40.04 | file | lirakis, what is your config? |
| 19:40.11 | mjordan | but that won't "fix" everything, nor should it - it should assert |
| 19:40.13 | opticron | actually, it's not a null instance |
| 19:40.18 | mjordan | er |
| 19:40.26 | mjordan | what is the RTP Engine callback? |
| 19:40.47 | opticron | ast_rtcp_write_report |
| 19:41.13 | mjordan | whaa.. |
| 19:41.18 | file | uh |
| 19:41.25 | opticron | errr |
| 19:41.37 | lirakis | file this is the peer in sip.conf http://pastie.org/private/km8e0jxh91ya1dx31edlpg |
| 19:41.51 | mjordan | opticron: bt? |
| 19:42.02 | opticron | it's running under the scheduler |
| 19:42.04 | lirakis | i just checked the sdp for 1 version older of chrome - it is sending a=crypto lines, while new versions are not |
| 19:42.07 | file | lirakis, there's no DTLS-SRTP configuration there |
| 19:42.24 | mjordan | opticron: right, but what line is it attempting to get a format on? That's RTCP |
| 19:42.29 | mjordan | it shouldn't give a fig about formats |
| 19:42.40 | opticron | 2633 |
| 19:42.50 | file | lirakis, http://pastebin.com/bzUPbR9q is a config that works and there is a script in contrib/scripts called ast_tls_cert which can be used to produce the required cert |
| 19:42.54 | mjordan | k, that's a cached frame on the RTP instance |
| 19:43.04 | opticron | gettign rate for jitter |
| 19:43.06 | mjordan | why is that format NULL? |
| 19:43.13 | lirakis | file, i guess... things were perhaps just working with SRTP |
| 19:43.22 | lirakis | and now that they are dropping it we are realizing it never worked with dtls |
| 19:43.23 | lirakis | heh |
| 19:43.38 | opticron | wonders if the frame was destroyed or otherwise cached improperly |
| 19:44.07 | mjordan | opticron: the payload's format is ao2_replace'd on that frame on each read |
| 19:44.17 | file | lirakis, and that's why I spent a week working on it >_> |
| 19:44.21 | lirakis | file that makes more sense |
| 19:44.32 | file | silly browsers |
| 19:44.54 | mjordan | opticron: what I don't know is why the payload's format would ever be NULL |
| 19:45.17 | lirakis | yeah - sorry i didnt do any of our initial set up for this stuff - i guess the dev who did - didnt come to that realization.... or look at the dang code |
| 19:45.18 | lirakis | hehe |
| 19:45.54 | lirakis | pretty ovbious looking at like 10535 of chan_sip.c |
| 19:46.11 | *** join/#asterisk-dev newtonr (~newtonr@nat/digium/x-gtgqrpauldeuccrh) |
| 19:46.12 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 19:46.15 | mjordan | opticron: I think you should assert on line 4166 on payload->format |
| 19:46.18 | lirakis | file, thanks again for your help - and patience |
| 19:46.41 | mjordan | opticron: you may also want to determine if this is happening on shutdown of the RTP instance, which is quite possible |
| 19:47.04 | mjordan | opticron: Another approach you could take would be for the cached format on the frame to never be ref bumped, and have it hold just a pointer to a format. That would be a bit more "risky" however. |
| 19:48.30 | mjordan | opticron: by the time we reach the point in the code where we are doing that ao2_replace, we should have a valid payload with an Asterisk format... unless there's something very wrong going on |
| 19:48.48 | opticron | I'll look further into it |
| 19:56.23 | lirakis | im guessing we did this before - and the guide just was flubbed |
| 20:02.34 | jrose_atDigium | mjordan + correyfarrell -- I've posted a review for some of the chan_iax2 work I'm doing in formats. Haven't moved any functions yet, but I do have chan_iax2 running again, so that's something. |
| 20:02.49 | jrose_atDigium | Also still haven't issues with codec preferences. |
| 20:02.55 | jrose_atDigium | *still having |
| 20:10.03 | opticron | cranks up valgrind with --track-origins=yes |
| 20:13.08 | *** part/#asterisk-dev lirakis (~lirakis@50-198-204-238-static.hfc.comcastbusiness.net) |
| 20:13.09 | mjordan | opticron: I may be hitting the same crash you are |
| 20:15.23 | coreyfarrell | mjordan / opticron: you have a backtrace? |
| 20:17.31 | opticron | 1 sec |
| 20:17.38 | opticron | I can generate it again |
| 20:17.47 | mjordan | coreyfarrell: Yes, but it's pretty useless on my end. We have an off by one clean up of a format. |
| 20:18.26 | opticron | mjordan, have you applied https://reviewboard.asterisk.org/r/3746/ ? |
| 20:19.21 | coreyfarrell | so maybe the refs log would be useful too |
| 20:20.03 | opticron | http://pastebin.com/SxQqJebU |
| 20:20.12 | coreyfarrell | looks |
| 20:20.22 | opticron | mine doesn't appear to be a ref issue, just something getting cleared that shouldn't be |
| 20:24.48 | coreyfarrell | opticron: what channel driver is this? any chance you want to print *rtp? |
| 20:25.22 | opticron | chan_pjsip |
| 20:25.23 | opticron | I can |
| 20:26.12 | opticron | http://pastebin.com/AXixpews |
| 20:26.56 | opticron | everthing is zero... |
| 20:30.30 | mjordan | not everything |
| 20:30.47 | mjordan | which rtp instance is it? |
| 20:30.51 | mjordan | your audio or your video instance? |
| 20:31.29 | mjordan | and if we haven't gotten anything yet to set the payload ... that would be bad |
| 20:32.50 | coreyfarrell | I'm scanning res_rtp_asterisk for all instances of the whole word 'f'. one thing I just saw that probably isn't related, process_dtmf_cisco() calls create_dtmf_frame() and doesn't duplicate the frame. everything else that uses create_dtmf_frame calls ast_frdup on the result. |
| 20:36.09 | mjordan | nothing calls process_dtmf_cisco |
| 20:36.18 | mjordan | really, at any rate. |
| 20:36.33 | mjordan | There's two other things I noticed, which is that we can memset or overwrite a format on the static list |
| 20:36.42 | mjordan | but that's not going to cause the errors I'm seeing. |
| 20:37.54 | mjordan | we really need to use tags more. |
| 20:38.06 | mjordan | It's kind of impossible with this many reference changes to know what is supposed to line up with what |
| 20:38.31 | opticron | I've gotten decent at guessing |
| 20:38.38 | opticron | but yes, tags |
| 20:39.58 | mjordan | so, here's a rather radical thought. |
| 20:40.09 | mjordan | What if most things didn't change the ref count of a format. |
| 20:40.17 | mjordan | including frames. |
| 20:40.24 | coreyfarrell | if someone wants to upload a refs log I can try searching it |
| 20:40.33 | mjordan | sure, give me a sec |
| 20:40.34 | coreyfarrell | please not pastebin :) |
| 20:40.42 | file | mjordan, continue... |
| 20:40.59 | mjordan | coreyfarrell: I'll e-mail mine to you |
| 20:44.22 | mjordan | well, formats, once created, are not destroyed (or better not be) until asterisk shutdown |
| 20:44.29 | mjordan | by shutdown, I mean long after all dynamic modules have unloaded |
| 20:44.47 | mjordan | in fact, the things that provide formats (codecs) cannot be unloaded |
| 20:44.58 | mjordan | and can only be disposed of just moments before the process exits |
| 20:45.16 | mjordan | essentially, the vast majority of ref counting that we are doing is pointless: the lifetime of the object is relatively well defined |
| 20:45.32 | mjordan | the exception, of course, are new formats created as a result of attributes. Some a format cap structure does need to increment the ref count |
| 20:45.33 | file | WELL |
| 20:45.35 | file | yeah |
| 20:45.36 | coreyfarrell | mjordan: that is the case for cached formats.. |
| 20:45.38 | mjordan | as it is a container that holds the ref attribute |
| 20:45.57 | mjordan | however, most of those containers also have a fixed lifetime: typically, the lifetime of the channel |
| 20:46.03 | mjordan | shrugs |
| 20:46.16 | file | mjordan, I think... that would... work |
| 20:46.35 | mjordan | possibly. If a frame lasts beyond the lifetime of a channel, then it needs to not touch the format |
| 20:46.53 | mjordan | I'm not sure that a frame can last beyond the lifetime of a channel in any way that it would be used |
| 20:48.02 | mjordan | I'm not sure I love this idea, but looking at our ref logs right now - things don't look very promising |
| 20:48.06 | *** join/#asterisk-dev sgriepentrog (~sgriepent@107.220.72.78) |
| 20:48.50 | file | traded copying a chunk of memory a lot for ref logs of insanity |
| 20:49.21 | mjordan | yeah |
| 20:49.37 | mjordan | I'm hesitant to change much just yet |
| 20:49.54 | mjordan | I'm going to try tagging some of the more common places in my ref log and see if that doesn't narrow some of these down |
| 20:52.31 | coreyfarrell | mjordan: I think r3743 can help in some cases (ast_rtp_read line 4167 / 4168 for example) |
| 20:52.52 | mjordan | k |
| 20:53.26 | mjordan | coreyfarrell: I'm going to post a tag variant of ao2_bump/ao2_replace/ao2_cleanup |
| 20:53.49 | coreyfarrell | sounds good |
| 21:06.11 | coreyfarrell | anyone object if I commit r3743 now? This is for 12/trunk, but ao2_replace isn't actually used at all in 12/trunk. |
| 21:06.55 | mjordan | nope, go for it |
| 21:10.59 | opticron | 3750 needs an update for 3743 having gone in |
| 21:11.46 | coreyfarrell | was just going to say that :) |
| 21:13.23 | mjordan | yup, will do |
| 21:14.18 | coreyfarrell | mjordan: also when you get a chance take a look at my last comment on 3734 about checking cap counts |
| 21:23.53 | *** join/#asterisk-dev ipengineer_ (~zconkle@static-71-252-134-63.dllstx.fios.verizon.net) |
| 21:33.05 | *** join/#asterisk-dev otherwiseguy (~otherwise@nat/cisco/x-chibeiqfibijsxiz) |
| 22:00.27 | coreyfarrell | mjordan: 3751 applies to the ref's log you emailed me. |
| 22:00.56 | coreyfarrell | can't say it's the only cause, but your log shows fmt being unref'ed at both lines |
| 22:11.29 | mjordan | blech. I hate it when code I write ends up being the culprit :-P |
| 22:12.44 | mjordan | I missed the later ao2_ref(fmt, -1) |
| 22:13.41 | mjordan | opticron: since that's in sip_new, it's probably affecting you as well |
| 22:13.45 | mjordan | er, not opticron |
| 22:13.47 | mjordan | sgriepentrog |
| 22:13.49 | *** join/#asterisk-dev ipengineer (~zconkle@static-71-252-134-63.dllstx.fios.verizon.net) |
| 22:14.28 | opticron | :) |
| 22:14.31 | opticron | I'd hope not |
| 22:24.37 | mjordan | compiles |
| 22:27.16 | *** join/#asterisk-dev ipengineer (~zconkle@static-71-252-134-63.dllstx.fios.verizon.net) |
| 22:38.49 | *** join/#asterisk-dev otherwiseguy (~otherwise@72.21.225.66) |
| 22:41.53 | *** join/#asterisk-dev mjordan (~mjordan@nat/digium/x-fzlwcjqjvrorodqj) |
| 22:41.53 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 22:42.03 | mjordan | coreyfarrell: woot, that was it |
| 22:43.04 | *** part/#asterisk-dev kharwell (~kharwell@nat/digium/x-uzmfwcsmdwimmiqy) |
| 22:57.18 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@64.235.216.2) |
| 23:27.42 | *** join/#asterisk-dev otherwiseguy (~otherwise@72.21.225.66) |
