Bug#921266: The segfault Is defintely not caused by t-online, but linpone
Am 09.04.19 um 16:20 schrieb Jonas Smedegaard: > Quoting Alf (2019-04-08 22:08:27) >> I have done further tests and finally given up to get linphone working >> as a SIP client here - the segfault on password enty is absolutely >> reproducible. >> >> To solve this issue I have tried "twinkle" (another SIP client from >> Buster-repo and that was set up within a few minutes. >> Just entered the SIP and STUN servers and my credentials for my >> t-online account - and all works as expected. >> >> This definitely confirms that it is not related to t-online, nor to >> some missconfiguration of my system. >> >> Moreover such a segfault is caused by severe bugs in the application >> itself. Again the syslog message: >> >> tux kernel: [ 4866.131680] linphone[6107]: segfault at 0 ip >> 003153a8f37a sp 7ffe572604c8 error 4 in >> libc-2.24.so[3153a0+194000] >> Mar 3 18:17:33 tux kernel: [ 4866.131690] Code: 89 e6 5b 4c 89 ef 5d 41 >> 5c 41 5d e9 e0 49 ff ff 89 f8 31 d2 66 0f ef ff 09 f0 25 ff 0f 00 00 3d >> c0 0f 00 00 0f 8f 78 02 00 00 0f 6f 0f f3 0f 6f 06 66 0f 74 c1 66 >> 0f da c1 66 0f ef c9 66 0f >> >> of which the tail is always the same: >> - >> ... 89 f8 31 d2 66 0f ef ff 09 f0 25 ff 0f 00 00 3d c0 0f 00 00 0f 8f 78 >> 02 00 00 0f 6f 0f f3 0f 6f 06 66 0f 74 c1 66 0f da c1 66 0f ef c9 66 0f >> >> There is also no means to gather more info about the crash in libc6 >> besides the syslog entry. Linphone's debug output. >> Even at highest level does not disclose anything valuable for me. >> >> So linphone 3.12 from Buster is buggy and absolutely unusable for me, >> severity should be set back to 'serious' to reflect that. > > Above crash involves libc-2.24 which is not Buster, so it seems you test > in some mixture of stable and testing system. > > Probably helpful to know which official Debian package repository you > primarily use, and which packages deviate from that - in any way, > including held packages and packages from deb-multimedia.org. > > Please also mention code included by other means than official Debian > packages - e.g. self-compiled kernel, flatpak... > > > - Jonas > Sorry, don't know how that happened - shame on me. Of course Buster has: ii libc6:amd642.28-8 amd64GNU C Library: Shared libraries Here the segfault message generated a few minutes ago in Buster. You can see, also here the tail is identical Apr 9 17:54:37 tux kernel: [ 203.438238] linphonec[1581]: segfault at 0 ip 7fd44a1ecc7a sp 7ffcb9a79018 error 4 in libc-2.28.so[7fd44a17e000+148000] Apr 9 17:54:37 tux kernel: [ 203.438248] Code: 0f b6 0e 0f b6 07 29 c8 c3 0f 1f 80 00 00 00 00 89 f8 31 d2 66 0f ef ff 09 f0 25 ff 0f 00 00 3d c0 0f 00 00 0f 8f 78 02 00 00 0f 6f 0f f3 0f 6f 06 66 0f 74 c1 66 0f da c1 66 0f ef c9 66 0f
Bug#921266: The segfault Is defintely not caused by t-online, but linpone
Am 09.04.19 um 14:54 schrieb W. Martin Borgert: > Quoting Alf : >> This definitely confirms that it is not related to t-online, nor to some >> missconfiguration of my system. > > Still, it would be interesting what leads to that behaviour. > I use Linphone as my one and only telephone and it works fine > for me on both a stretch and a buster system, both amd64. No > crashes at all. I use different providers, however. > I found (hopefully) an explanation how authentication at t-online should work. See this URL: https://andrewjprokop.wordpress.com/2015/01/27/understanding-sip-authentication/ If you have a look at my debug output some posts above you see that for registration the client contacts the server. The server responds with: SIP/2.0 401 Unauthorized 11030230348 and next it sends the nonce WWW-Authenticate: Digest realm="tel.t-online.de",nonce="415F6A393573AB5CED81AB3F",stale=true,algorithm=MD5,qop="auth" and so on until registration is complete This cycle is run twice with linphone 3.6.1 until registration is complete. It can be seen in the debug output id you search for the string "Authentication-Info: qop=auth" Linphone 3.12.0 only runs this cycle once and then crashing at following output: Authentication-Info: qop=auth,rspauth="3f09707f2c9d40af03c39831078ffc75",cnonce="zJy0v3v2zDA0WMNm",nc=0001 2019-04-08 18:13:30:008 ortp-message-channel [0x55623f893e60] [722] bytes parsed 2019-04-08 18:13:30:008 ortp-message-Found transaction matching response. 2019-04-08 18:13:30:008 ortp-message-Updating auth context for ream [tel.t-online.de] next nonce is going to be [(null)] crash! It for sure has nothing to do with the GUI, all debug output was collected using CLI vith option "-d 5". One final thought: Ubuntu/Canonical has choosen linphone 3.6.1 for the upcoming release 1904 (Disco Dingo) instead of 3.12 - maybe they have a reason for that decision, expecially as they genally aim for latest available version?
Bug#921266: The segfault Is defintely not caused by t-online, but linpone
Quoting Alf (2019-04-08 22:08:27) > I have done further tests and finally given up to get linphone working > as a SIP client here - the segfault on password enty is absolutely > reproducible. > > To solve this issue I have tried "twinkle" (another SIP client from > Buster-repo and that was set up within a few minutes. > Just entered the SIP and STUN servers and my credentials for my > t-online account - and all works as expected. > > This definitely confirms that it is not related to t-online, nor to > some missconfiguration of my system. > > Moreover such a segfault is caused by severe bugs in the application > itself. Again the syslog message: > > tux kernel: [ 4866.131680] linphone[6107]: segfault at 0 ip > 003153a8f37a sp 7ffe572604c8 error 4 in > libc-2.24.so[3153a0+194000] > Mar 3 18:17:33 tux kernel: [ 4866.131690] Code: 89 e6 5b 4c 89 ef 5d 41 > 5c 41 5d e9 e0 49 ff ff 89 f8 31 d2 66 0f ef ff 09 f0 25 ff 0f 00 00 3d > c0 0f 00 00 0f 8f 78 02 00 00 0f 6f 0f f3 0f 6f 06 66 0f 74 c1 66 > 0f da c1 66 0f ef c9 66 0f > > of which the tail is always the same: > - > ... 89 f8 31 d2 66 0f ef ff 09 f0 25 ff 0f 00 00 3d c0 0f 00 00 0f 8f 78 > 02 00 00 0f 6f 0f f3 0f 6f 06 66 0f 74 c1 66 0f da c1 66 0f ef c9 66 0f > > There is also no means to gather more info about the crash in libc6 > besides the syslog entry. Linphone's debug output. > Even at highest level does not disclose anything valuable for me. > > So linphone 3.12 from Buster is buggy and absolutely unusable for me, > severity should be set back to 'serious' to reflect that. Above crash involves libc-2.24 which is not Buster, so it seems you test in some mixture of stable and testing system. Probably helpful to know which official Debian package repository you primarily use, and which packages deviate from that - in any way, including held packages and packages from deb-multimedia.org. Please also mention code included by other means than official Debian packages - e.g. self-compiled kernel, flatpak... - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private signature.asc Description: signature
Bug#921266: The segfault Is defintely not caused by t-online, but linpone
Quoting Alf : This definitely confirms that it is not related to t-online, nor to some missconfiguration of my system. Still, it would be interesting what leads to that behaviour. I use Linphone as my one and only telephone and it works fine for me on both a stretch and a buster system, both amd64. No crashes at all. I use different providers, however.
Bug#921266: The segfault Is defintely not caused by t-online, but linpone
I have done further tests and finally given up to get linphone working as a SIP client here - the segfault on password enty is absolutely reproducible. To solve this issue I have tried "twinkle" (another SIP client from Buster-repo and that was set up within a few minutes. Just entered the SIP and STUN servers and my credentials for my t-online account - and all works as expected. This definitely confirms that it is not related to t-online, nor to some missconfiguration of my system. Moreover such a segfault is caused by severe bugs in the application itself. Again the syslog message: tux kernel: [ 4866.131680] linphone[6107]: segfault at 0 ip 003153a8f37a sp 7ffe572604c8 error 4 in libc-2.24.so[3153a0+194000] Mar 3 18:17:33 tux kernel: [ 4866.131690] Code: 89 e6 5b 4c 89 ef 5d 41 5c 41 5d e9 e0 49 ff ff 89 f8 31 d2 66 0f ef ff 09 f0 25 ff 0f 00 00 3d c0 0f 00 00 0f 8f 78 02 00 00 0f 6f 0f f3 0f 6f 06 66 0f 74 c1 66 0f da c1 66 0f ef c9 66 0f of which the tail is always the same: - ... 89 f8 31 d2 66 0f ef ff 09 f0 25 ff 0f 00 00 3d c0 0f 00 00 0f 8f 78 02 00 00 0f 6f 0f f3 0f 6f 06 66 0f 74 c1 66 0f da c1 66 0f ef c9 66 0f There is also no means to gather more info about the crash in libc6 besides the syslog entry. Linphone's debug output. Even at highest level does not disclose anything valuable for me. So linphone 3.12 from Buster is buggy and absolutely unusable for me, severity should be set back to 'serious' to reflect that.
