#2220 Segmentation fault on reset with the p51-d

[eatdirt]

Snapshot next of today, I have a segfault on reset confirmed with p51d, mirage-2000-5, SpaceShuttle etc... In fact, I think I have it with all planes, but it is hard to try them all :)

Nothing in the console, running linux.

[eatdirt]

Forgot to say, I am running all this with nasal-gc-threaded=false

[James Turner]

No back-trace for these I guess?

[eatdirt]

I'll try to get one!

[eatdirt]

Thread 9 "fgfs" received signal SIG32, Real-time event 32.

(gdb) bt
#0  0x00007f7fef84faf5 in pthread_cond_wait@@GLIBC_2.3.2 () from /lib64/libpthread.so.0
#1  0x00007f7fefa2c7ee in OpenThreads::Condition::wait(OpenThreads::Mutex*) () from /lib64/libOpenThreads.so.21
#2  0x0000000000c8752c in FGVoiceMgr::FGVoiceThread::wait_for_jobs (this=0x1c21b9b0)
    at /usr/src/debug/flightgear-2019.2.0-26.mga7.x86_64/src/Sound/voice.hxx:96
#3  FGVoiceMgr::FGVoiceThread::run (this=0x1c21b9b0) at /usr/src/debug/flightgear-2019.2.0-26.mga7.x86_64/src/Sound/voice.cxx:286
#4  0x00007f7fefa2c2e5 in OpenThreads::ThreadPrivateActions::StartThread(void*) () from /lib64/libOpenThreads.so.21
#5  0x00007f7fef84a04c in start_thread () from /lib64/libpthread.so.0
#6  0x00007f7fed2143af in clone () from /lib64/libc.so.6

[Stuart Buchanan]

All those are JSBSim aircraft. Are you getting the same problem on YASim?

[eatdirt]

Ok, I was trying to test a reset with the mirage-2000, yasim. Did not even reached that state, I got a segfault while flying it from OSG:

1. 467.83 [WARN]:OSG Error VertexBufferObject::removeBufferData(57251) out of range.
   467.83 [WARN]:OSG Error VertexBufferObject::removeBufferData(57191) out of range.
   467.83 [WARN]:OSG Error VertexBufferObject::removeBufferData(56823) out of range.
   467.83 [WARN]:OSG Error VertexBufferObject::removeBufferData(56808) out of range.
   467.83 [WARN]:OSG Error VertexBufferObject::removeBufferData(56958) out of range.
   467.83 [WARN]:OSG Error VertexBufferObject::removeBufferData(57191) out of range.
   467.83 [WARN]:OSG Error VertexBufferObject::removeBufferData(56937) out of range.
   Segmentation fault (core dumped)

I'll try to reproduce with a bt in another ticket.

[eatdirt]

All right, with mirage-2000, upon resetting, exactly same segfault:

Thread 8 "fgfs" received signal SIG32, Real-time event 32.
[Switching to Thread 0x7f711bb56700 (LWP 9812)]
0x00007f7139830af5 in pthread_cond_wait@@GLIBC_2.3.2 () from /lib64/libpthread.so.0
(gdb) bt
#0  0x00007f7139830af5 in pthread_cond_wait@@GLIBC_2.3.2 () from /lib64/libpthread.so.0
#1  0x00007f7139a0d7ee in OpenThreads::Condition::wait(OpenThreads::Mutex*) () from /lib64/libOpenThreads.so.21
#2  0x0000000000c974cc in FGVoiceMgr::FGVoiceThread::wait_for_jobs (this=0x16f0e0c0)
    at /usr/src/debug/flightgear-2020.2.0-2.mga7.x86_64/src/Sound/voice.hxx:96
#3  FGVoiceMgr::FGVoiceThread::run (this=0x16f0e0c0) at /usr/src/debug/flightgear-2020.2.0-2.mga7.x86_64/src/Sound/voice.cxx:286
#4  0x00007f7139a0d2e5 in OpenThreads::ThreadPrivateActions::StartThread(void*) () from /lib64/libOpenThreads.so.21
#5  0x00007f713982b04c in start_thread () from /lib64/libpthread.so.0
#6  0x00007f71372763af in clone () from /lib64/libc.so.6

[xDraconian]

SIG32 is a normal threading interrupt. That's masking the real problem.
Instruct GDB to ignore SIG32 via...
handle 32 pass nostop noprint

[James Turner]

Please re-test on next with commits 297e5095762be2fc0fe7995f7115e2e58edb09ae (SimGear) and 06e6883396fef04caf8e6c54d1874de2b5581644 (flightgear) - will hopefully make a difference.

It won't help the FGVoice one, but the removeBufferData one, I believe I've fixed.

[Gijs]

@dirteat. can we close this one?