New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Don't restore audio device during exporting #4083

Merged
merged 2 commits into from Mar 2, 2018

Conversation

Projects
None yet
2 participants
@PhysSong
Member

PhysSong commented Dec 30, 2017

This PR fixes deadlock on multi-track export with SDL.
Fixes #4032.

Don't restore audio device during exporting
Fixes deadlock on multi-track export with SDL

@PhysSong PhysSong requested review from zonkmachine and removed request for zonkmachine Dec 30, 2017

@@ -630,6 +627,17 @@ void Mixer::setAudioDevice( AudioDevice * _dev,
void Mixer::storeAudioDevice()
{
if( !m_oldAudioDev )

This comment has been minimized.

@PhysSong

PhysSong Jan 10, 2018

Member

I've included this check, but looks redundant. Maybe removing is better?

@@ -103,8 +102,6 @@ ProjectRenderer::ProjectRenderer( const Mixer::qualitySettings & qualitySettings
ProjectRenderer::~ProjectRenderer()
{
Engine::mixer()->restoreAudioDevice(); // also deletes audio-dev
Engine::mixer()->changeQuality( m_oldQualitySettings );

This comment has been minimized.

@zonkmachine

zonkmachine Jan 11, 2018

Member

Make sure #3680 still works. If these test projects renders, as instructed, it's a 'go' from me.

This comment has been minimized.

@PhysSong

PhysSong Jan 11, 2018

Member

@zonkmachine I'm pretty sure #3680 will still work because RenderManager destructor is called after ProjectRenderer destructor. However, testing is always good. I'll double-check it!

This comment has been minimized.

@PhysSong

PhysSong Jan 17, 2018

Member

@zonkmachine Tested. It still works!

This comment has been minimized.

@zonkmachine

zonkmachine Jan 17, 2018

Member

OK. I think you should merge.

@zonkmachine

Tested, works!
I don't have insight enough to review the code. It looks mighty fine though.

@zonkmachine

This comment has been minimized.

Member

zonkmachine commented Feb 6, 2018

Merge?

@zonkmachine

This comment has been minimized.

Member

zonkmachine commented Feb 25, 2018

I've managed to get it hung on the last step but can't replicate it. I don't know if I had #4189 cherry-picked on top when this happened or not. I think so but it seem to work fine generally, with or without it.

[Thread 0x7fffb1b10700 (LWP 15233) exited]
[New Thread 0x7fffb1b10700 (LWP 15309)]
[Thread 0x7fffb1b10700 (LWP 15309) exited]
^C
Program received signal SIGINT, Interrupt.
__lll_lock_wait () at ../nptl/sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:135
135	../nptl/sysdeps/unix/sysv/linux/x86_64/lowlevellock.S: No such file or directory.
(gdb) bt
#0  __lll_lock_wait () at ../nptl/sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:135
#1  0x00007ffff7bc6664 in _L_lock_952 () from /lib/x86_64-linux-gnu/libpthread.so.0
#2  0x00007ffff7bc64c6 in __GI___pthread_mutex_lock (mutex=0xd80320) at ../nptl/pthread_mutex_lock.c:114
#3  0x00007ffff62ab5fe in SDL_mutexP (mutex=<optimized out>) at ./src/thread/pthread/SDL_sysmutex.c:108
#4  0x000000000058265c in AudioSdl::stopProcessing (this=0xe7a5e0) at /home/zonkmachine/builds/lmms/lmms/src/core/audio/AudioSdl.cpp:110
#5  0x0000000000533fe3 in Mixer::stopProcessing (this=0xc45210) at /home/zonkmachine/builds/lmms/lmms/src/core/Mixer.cpp:251
#6  0x0000000000534f0a in Mixer::changeQuality (this=0xc45210, _qs=...) at /home/zonkmachine/builds/lmms/lmms/src/core/Mixer.cpp:564
#7  0x0000000000552036 in RenderManager::~RenderManager (this=0x189c050, __in_chrg=<optimized out>)
    at /home/zonkmachine/builds/lmms/lmms/src/core/RenderManager.cpp:54
#8  0x00000000005520a6 in RenderManager::~RenderManager (this=0x189c050, __in_chrg=<optimized out>)
    at /home/zonkmachine/builds/lmms/lmms/src/core/RenderManager.cpp:55
#9  0x000000000059b4a9 in ExportProjectDialog::accept (this=0x7fffffffcc30) at /home/zonkmachine/builds/lmms/lmms/src/gui/ExportProjectDialog.cpp:114
#10 0x00000000006438c2 in ExportProjectDialog::qt_static_metacall (_o=0x7fffffffcc30, _c=QMetaObject::InvokeMetaMethod, _id=2, _a=0x7fffffffc400)
    at /home/zonkmachine/builds/lmms/lmms/build/src/moc_ExportProjectDialog.cpp:56
#11 0x00007ffff697087a in QMetaObject::activate (sender=0x189c050, m=<optimized out>, local_signal_index=<optimized out>, argv=0x0)
    at kernel/qobject.cpp:3539
#12 0x000000000064ba9b in RenderManager::finished (this=0x189c050) at /home/zonkmachine/builds/lmms/lmms/build/src/moc_RenderManager.cpp:113
#13 0x00000000005521a1 in RenderManager::renderNextTrack (this=0x189c050) at /home/zonkmachine/builds/lmms/lmms/src/core/RenderManager.cpp:77
#14 0x000000000064b91e in RenderManager::qt_static_metacall (_o=0x189c050, _c=QMetaObject::InvokeMetaMethod, _id=2, _a=0x7fff9c0013c0)
    at /home/zonkmachine/builds/lmms/lmms/build/src/moc_RenderManager.cpp:57
@PhysSong

This comment has been minimized.

Member

PhysSong commented Feb 26, 2018

@zonkmachine What is the backtrace? When did you get that?

@zonkmachine

This comment has been minimized.

Member

zonkmachine commented Feb 26, 2018

@zonkmachine What is the backtrace? When did you get that?

The backtrace is from crashing/freezing at the end of rendering the project Ashore to separate tracks. I can't replicate this however and think I may have confused this branch and stable-1.2 as I would probably have tested the later to replicate the issue prior to testing the fix. The differences between the logs seem to be that this one is after the very last track is finished rendering and the one in #4032 (comment) is after rendering some two tracks or so.

I think we can ignore this backtrace and merge #4083.

@PhysSong

This comment has been minimized.

Member

PhysSong commented Feb 27, 2018

The backtrace seems to be from this PR. However, the problem looks like a separate issue and an edge case. I think I can fix that, but probably in another PR.

@zonkmachine

This comment has been minimized.

Member

zonkmachine commented Feb 27, 2018

I think I can fix that, but probably in another PR.

OK. Merge?

@PhysSong

This comment has been minimized.

Member

PhysSong commented Mar 2, 2018

OK. Merge?

I'll do that soon once Travis-CI build finishes.

@PhysSong PhysSong merged commit 926b654 into LMMS:stable-1.2 Mar 2, 2018

1 check passed

continuous-integration/travis-ci/pr The Travis CI build passed
Details

@PhysSong PhysSong deleted the PhysSong:multiexport branch Mar 2, 2018

@PhysSong PhysSong referenced this pull request Jun 6, 2018

Closed

SDL deadlock on exit #4406

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment