Re: [mythtv] Re: Re: SIGSEGV at end of recording
Charlie Brej wrote: Isaac Richards wrote: And, after talking with Chris about mythweb a bit, I think that last commit of mine may have fixed the issue. Hopefully. Maybe? Isaac I gave it a good hammering doing the things that normally make it die and by now I would have expected it to crash but it hasn't. I will give it another go tomorrow but I think that has fixed it. Thanks for hunting this one down. ___ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev I've been away for a while, but I've finally had a chance to test these fixes. I'm no longer able to crash the backend with any of the usual things that would cause problems, and it's been running like a charm. Thanks for looking into this. layttan ___ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
Re: [mythtv] Re: Re: SIGSEGV at end of recording
Isaac Richards wrote: And, after talking with Chris about mythweb a bit, I think that last commit of mine may have fixed the issue. Hopefully. Maybe? Isaac I gave it a good hammering doing the things that normally make it die and by now I would have expected it to crash but it hasn't. I will give it another go tomorrow but I think that has fixed it. Thanks for hunting this one down. -- Charlie Brej APT Group, Dept. of Computer Science, University of Manchester Web: http://brej.org/ Tel: +44 (0)161 275 6177 Mail: IT303, University of Manchester, Manchester, M13 9PL, UK ___ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
Re: [mythtv] Re: Re: SIGSEGV at end of recording
On Thursday 26 May 2005 07:14 pm, Isaac Richards wrote: > On Thursday 26 May 2005 11:11 am, Isaac Richards wrote: > > On Thursday 26 May 2005 04:24 am, Charlie Brej wrote: > > > Here is a dump with the current CVS. > > > I can often reproduce it when I browse with mythweb and then the next > > > recording finishes. > > > > Oh well, someone else will have to fix it. There's no way it should be > > dying at this point, unless your STL is broken (which is extremely > > unlikely). > > I did just go add some additional debugging statements, so I'd be > interested in seeing what happens now (still keeping the printfs you > added). Please include the full log of the run, not a trimmed version. And, after talking with Chris about mythweb a bit, I think that last commit of mine may have fixed the issue. Hopefully. Maybe? Isaac ___ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
Re: [mythtv] Re: Re: SIGSEGV at end of recording
On Thursday 26 May 2005 11:11 am, Isaac Richards wrote: > On Thursday 26 May 2005 04:24 am, Charlie Brej wrote: > > Here is a dump with the current CVS. > > I can often reproduce it when I browse with mythweb and then the next > > recording finishes. > > Oh well, someone else will have to fix it. There's no way it should be > dying at this point, unless your STL is broken (which is extremely > unlikely). I did just go add some additional debugging statements, so I'd be interested in seeing what happens now (still keeping the printfs you added). Please include the full log of the run, not a trimmed version. Isaac ___ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
Re: [mythtv] Re: Re: SIGSEGV at end of recording
On Thursday 26 May 2005 04:24 am, Charlie Brej wrote: > Isaac Richards wrote: > > On Wednesday 25 May 2005 04:24 pm, Charlie Brej wrote: > >>The original guy didn't reply yet but I have the same problem. I inserted > >>the lines above and here is the dump. > > > > Ok. Please try to reproduce with current CVS. > > > > Isaac > > ___ > > mythtv-dev mailing list > > mythtv-dev@mythtv.org > > http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev > > Here is a dump with the current CVS. > I can often reproduce it when I browse with mythweb and then the next > recording finishes. Oh well, someone else will have to fix it. There's no way it should be dying at this point, unless your STL is broken (which is extremely unlikely). Isaac ___ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
Re: [mythtv] Re: Re: SIGSEGV at end of recording
Isaac Richards wrote: On Wednesday 25 May 2005 04:24 pm, Charlie Brej wrote: The original guy didn't reply yet but I have the same problem. I inserted the lines above and here is the dump. Ok. Please try to reproduce with current CVS. Isaac ___ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev Here is a dump with the current CVS. I can often reproduce it when I browse with mythweb and then the next recording finishes. -8<- Two Pints of Lager and a Packet of 7 1002 31 02:00-02:30 1 0 0 A P 2 The Mark Steel Lecture - "Ludwig va 10 1020 31 03:25-03:55 1 0 0 A P 2 Foster's Home for Imaginary Friends 36 1017 31 16:00-17:00 1 2 2 A 2 2 The Simpsons - "Marge in Chains" 4 1009 31 18:00-18:30 1 0 0 A P 2 The Real Good Life - "1/7" 3 1006 31 20:00-21:00 1 2 2 S 2 2 Scrapheap Challenge - "Fire Engine" 36 1017 31 21:00-22:00 1 2 2 A 2 2 CSI: Miami - "Tinder Box" 5 1011 31 22:00-22:55 1 2 2 A 2 2 Two Pints of Lager and a Packet of 7 1002 31 23:00-23:30 1 0 0 A E 2 101 Best Kept Hollywood Secrets - "5 1011 31 23:25-00:30 1 2 2 A 2 2 Two Pints of Lager and a Packet of 7 1002 01 01:30-02:00 1 0 0 A P 2 Foster's Home for Imaginary Friends 36 1017 01 16:00-17:00 1 2 2 A 2 2 The Simpsons - "Krusty Gets Cancell4 1009 01 18:00-18:30 1 0 0 A P 2 Holidays in the Danger Zone: Places2 1001 01 19:30-20:00 1 2 2 A 2 2 Super Structures - "International S 36 1017 01 21:00-22:00 1 2 2 A 2 2 Foster's Home for Imaginary Friends 36 1017 02 16:00-17:00 1 2 2 A 2 2 The Simpsons - "Homer's Barbershop 4 1009 02 18:00-18:30 1 0 0 A P 2 Dead Ringers 2 1001 02 21:00-21:30 1 2 2 A 2 2 The Thick of It 10 1020 02 22:30-23:00 1 2 2 A 2 2 The Thick of It 10 1020 02 23:45-00:15 1 2 2 A 2 2 The Thick of It 10 1020 03 01:45-02:15 1 2 2 A 2 2 Foster's Home for Imaginary Friends 36 1017 03 16:00-17:00 1 2 2 A 2 2 The Simpsons - "Cape Feare"4 1009 03 18:00-18:30 1 0 0 A P 2 Have I Got News for You1 1000 03 21:00-21:30 1 2 2 A 2 2 Two Pints of Lager and a Packet of 7 1002 03 23:00-23:30 1 0 0 A P 2 Kerrang! Superstars of Noise 18 1030 03 23:00-00:00 1 2 2 A 2 2 Two Pints of Lager and a Packet of 7 1002 03 23:30-00:00 1 0 0 A P 2 Ever Wondered about Food - "Mushroo2 1001 04 11:30-12:00 1 2 2 A 2 2 Foster's Home for Imaginary Friends 36 1017 04 16:30-17:00 1 2 2 A 2 2 Doctor Who - "Boom Town" 1 1000 04 19:00-19:45 1 2 2 A 2 2 Doctor Who Confidential - "Unsung H7 1002 04 19:45-20:15 1 2 2 A 2 2 The Thick of It 10 1020 04 21:00-21:30 1 2 2 A 2 2 Three's Outtakes 7 1002 04 22:00-22:05 1 2 2 A 2 2 Law and Order: Criminal Intent - "G5 1011 04 22:25-23:20 1 2 2 A 2 2 Have I Got News for You2 1001 04 23:25-23:55 1 2 2 A 2 2 The Graduate 3 1006 04 23:25-01:20 1 1 1 F 1 -1 Dead Ringers 2 1001 04 23:55-00:25 1 2 2 A 2 2 Doctor Who - "Boom Town" 7 1002 05 00:20-01:05 1 0 0 A E 2 The Thick of It 10 1020 05 00:25-00:55 1 2 2 A 2 2 Doctor Who Confidential - "Unsung H7 1002 05 01:05-01:35 1 0 0 A E 2 Law and Order: Special Victims Unit5 1011 05 01:15-02:10 1 2 2 A 2 5 Two Pints of Lager and a Packet of 7 1002 05 01:35-02:05 1 0 0 A P 2 Two Pints of Lager and a Packet of 7 1002 05 02:05-02:35 1 0 0 A P 2 Foster's Home for Imaginary Friends 36 1017 05 16:30-17:00 1 2 2 A 2 2 Doctor Who - "Boom Town" 7 1002 05 19:00-19:45 1 0 0 A E 2 Doctor Who Confidential - "Unsung H7 1002 05 19:45-20:00 1 0 0 A E 2 Horizon - "Malaria: Defeating the C2 1001 05 21:00-21:50 1 2 2 A 2 2 Two Pints of Lager and a Packet of 7 1002 05 23:00-23:30 1 0 0 A P 2 Two Pints of Lager and a Packet of 7 1002 05 23:30-00:00 1 0 0 A P 2 Dexter's Laboratory 36 1017 06 12:30-13:00 1 2 2 A 2 2 Foster's Home for Imaginary Friends 36 1017 06 13:30-14:00 1 2 2 A 2 2 Foster's Home for Imaginary Friends 36 1017 06 16:30-17:00 1 2 2 A 2 2 The Simpsons - "Homer Goes to Colle4 1009 06 18:00-18:30 1 0 0 A P 2 ER - "The Show Must Go On" 4 1009 06 21:00-22:00 1 2 2 A 2 2 Two Pints of Lager and a Packet of 7 1002 06 21:30-22:00 1 0 0 A P 2 Tarrant on TV 3 1006 06 23:00-23:30 1 2 2 A 2 2 Men Behaving Badly - "Jealousy"1 1000 06 23:05-23:35 1 1 1 A 1 0 Two Pints of Lager and a Packet of 7 1002 07 02:00-02:
Re: [mythtv] Re: Re: SIGSEGV at end of recording
On Wednesday 25 May 2005 04:24 pm, Charlie Brej wrote: > The original guy didn't reply yet but I have the same problem. I inserted > the lines above and here is the dump. Ok. Please try to reproduce with current CVS. Isaac ___ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
Re: [mythtv] Re: Re: SIGSEGV at end of recording
> Can you put something like: > > printf("deleting refsock: %p\n", this); in the 'if (refCount < 0)' clause in server.h? > Also, something like: > printf("new socket: %p\n", this); in the constructor, and printf("locking socket %p\n", this); in the Lock + unlock functions (change text as appropriate, of course). > > Don't really understand how it could be crashing, really. Things shouldn't be able to be deleted and still stay in that list, unless.. This isn't a slave backend doing the dying, is it? > > Isaac The original guy didn't reply yet but I have the same problem. I inserted the lines above and here is the dump. Seems 0xab716f28 is deleted and then locked. ---8<--- locking socket 0xabafab20 unlocking socket 0xabafab20 locking socket 0xad586c20 2005-05-25 20:53:48.802 write -> 33 45 BACKEND_MESSAGE[]:[]SCHEDULE_CHANGE[]:[]empty unlocking socket 0xad586c20 locking socket 0xab716f28 unlocking socket 0xab716f28 locking socket 0xab854da0 unlocking socket 0xab854da0 2005-05-25 20:53:50.000 JobQueue::GetJobsInQueue: findJobs search bitmask 28, found 0 total jobs new socket: 0xaba6ded0 locking socket 0xaba6ded0 2005-05-25 20:53:59.504 write -> 22 13 ACCEPT[]:[]17 unlocking socket 0xaba6ded0 locking socket 0xaba6ded0 2005-05-25 20:53:59.523 read <- 22 26 ANN Playback hilly.house 0 2005-05-25 20:53:59.534 MainServer::HandleAnnounce Playback 2005-05-25 20:53:59.538 adding: hilly.house as a client (events: 0) 2005-05-25 20:53:59.538 write -> 22 2 OK unlocking socket 0xaba6ded0 locking socket 0xaba6ded0 2005-05-25 20:53:59.723 read <- 22 19 QUERY_GETALLPENDING 2005-05-25 20:53:59.738 write -> 22 70372 0[]:[]121[]:[]My Hero[]:[]Time and Time Again[]:[]Sitcom about an... unlocking socket 0xaba6ded0 deleting refsock: 0xad4dfb18 deleting refsock: 0xab716f28 deleting refsock: 0xab8c0fa0 new socket: 0xabb0b988 locking socket 0xabb0b988 2005-05-25 20:54:12.471 write -> 22 13 ACCEPT[]:[]17 unlocking socket 0xabb0b988 locking socket 0xabb0b988 2005-05-25 20:54:12.496 read <- 22 26 ANN Playback hilly.house 0 2005-05-25 20:54:12.502 MainServer::HandleAnnounce Playback 2005-05-25 20:54:12.502 adding: hilly.house as a client (events: 0) 2005-05-25 20:54:12.503 write -> 22 2 OK unlocking socket 0xabb0b988 locking socket 0xabb0b988 2005-05-25 20:54:12.761 read <- 22 19 QUERY_GETALLPENDING 2005-05-25 20:54:12.782 write -> 22 70372 0[]:[]121[]:[]My Hero[]:[]Time and Time Again[]:[]Sitcom about an... unlocking socket 0xabb0b988 new socket: 0xab716f28 locking socket 0xab716f28 2005-05-25 20:54:15.053 write -> 22 13 ACCEPT[]:[]17 unlocking socket 0xab716f28 locking socket 0xab716f28 2005-05-25 20:54:15.079 read <- 22 26 ANN Playback hilly.house 0 2005-05-25 20:54:15.090 MainServer::HandleAnnounce Playback 2005-05-25 20:54:15.090 adding: hilly.house as a client (events: 0) 2005-05-25 20:54:15.090 write -> 22 2 OK unlocking socket 0xab716f28 locking socket 0xab716f28 2005-05-25 20:54:15.256 read <- 22 19 QUERY_GETALLPENDING 2005-05-25 20:54:15.271 write -> 22 70372 0[]:[]121[]:[]My Hero[]:[]Time and Time Again[]:[]Sitcom about an... unlocking socket 0xab716f28 2005-05-25 20:54:15.370 JobQueue currently set at 1 job(s) max and to run new jobs from 00:00 to 23:59 2005-05-25 20:54:15.380 JobQueue::GetJobsInQueue: findJobs search bitmask 4, found 0 total jobs deleting refsock: 0xab6bd850 deleting refsock: 0xaba43530 deleting refsock: 0xab854da0 deleting refsock: 0xaba6ded0 2005-05-25 20:55:02.168 DVB#1 Trying to tune to channel 5. 2005-05-25 20:55:02.179 DVB#1 Frequency: 850166670. BW:8MHz HP:2/3 LP:1/2 C:64 TM:2K H:None GI:1/32 2005-05-25 20:55:02.241 Started recording "Super Structures" on channel: 1017 on cardid: 1, sourceid 1 2005-05-25 20:55:02.250 DVB#0 Trying to tune to channel 36. 2005-05-25 20:55:02.260 DVB#0 Frequency: 850166670. BW:8MHz HP:2/3 LP:1/2 C:64 TM:2K H:None GI:1/32 locking socket 0xabafab20 unlocking socket 0xabafab20 locking socket 0xad586c20 2005-05-25 20:55:02.328 write -> 33 45 BACKEND_MESSAGE[]:[]SCHEDULE_CHANGE[]:[]empty unlocking socket 0xad586c20 locking socket 0xab716f28 unlocking socket 0xab716f28 2005-05-25 20:55:02.446 DVB#0 Waiting for frontend event after tune. 2005-05-25 20:55:02.721 DVB#0 DVB signal cdcd | snr bdbd | ber 1fffe | unc0 2005-05-25 20:55:03.027 DVB#0 Status: LOCK. 2005-05-25 20:55:03.028 DVB#0 Multiplex Locked 2005-05-25 20:55:03.718 DVB#1 Waiting for frontend event after tune. 2005-05-25 20:55:03.742 DVB#1 DVB signal 3a3a | snr | ber0 | unc0 2005-05-25 20:55:03.742 DVB#1 Status: LOCK. 2005-05-25 20:55:03.743 DVB#1 Multiplex Locked 2005-05-25 20:55:03.919 DVB#0 Successfully tuned to channel 36. [New Thread -1390416976 (LWP 12346)] locking socket 0xabafab20 unlocking socket 0xabafab20 locking socket 0xad586c20 2005-05-25 20:55:04.085 write -> 33 51 BACKEND_MESSAGE[]:[]RECORDING_LIST_CHANGE[]:[]empty unlocki
Re: [mythtv] Re: Re: SIGSEGV at end of recording
On Sunday 22 May 2005 10:01 am, Michael J. Sherman wrote: > CVS pulled at 11:13am on May 21 (Saturday). I also got a similar crash > not after the recording finished, but after the commercial flagging > finished. But the crash was still with the QMutex::lock. Checking > back through the archives, it seems a few folks have been getting this > crash after recordings finish, after commercial flagging, and after > deleting via MythWeb. > > Below is the complete backtrace of a crash after a recording finished. Can you put something like: printf("deleting refsock: %p\n", this); in the 'if (refCount < 0)' clause in server.h? Also, something like: printf("new socket: %p\n", this); in the constructor, and printf("locking socket %p\n", this); in the Lock + unlock functions (change text as appropriate, of course). Don't really understand how it could be crashing, really. Things shouldn't be able to be deleted and still stay in that list, unless.. This isn't a slave backend doing the dying, is it? Isaac ___ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev