1861709 - Intermittent Assertion failure: fallback->InIteration(), at /builds/worker/checkouts/gecko/dom/media/GraphDriver.cpp:784

Status: RESOLVED FIXED

(Core :: Audio/Video: MediaStreamGraph, defect, P5)

Description

[Treeherder Bug Filer]

Filed by: ncsoregi [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=434050042&repo=autoland
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/NtBpXfvZTTyQZmXeGyPovQ/runs/0/artifacts/public/logs/live_backing.log

---

[task 2023-10-27T14:15:04.418Z] 14:15:04     INFO - TEST-START | dom/media/webrtc/tests/mochitests/test_getUserMedia_mediaStreamConstructors.html
[task 2023-10-27T14:15:04.496Z] 14:15:04     INFO - GECKO(3067) | TEST DEVICES: No test device found in media.audio_loopback_dev, using fake audio streams.
[task 2023-10-27T14:15:04.497Z] 14:15:04     INFO - GECKO(3067) | TEST DEVICES: No test device found in media.video_loopback_dev, using fake video streams.
[task 2023-10-27T14:15:04.525Z] 14:15:04     INFO - GECKO(3067) | Assertion failure: fallback->InIteration(), at /builds/worker/checkouts/gecko/dom/media/GraphDriver.cpp:784
[task 2023-10-27T14:15:04.529Z] 14:15:04     INFO -  Initializing stack-fixing for the first stack frame, this may take a while...
[task 2023-10-27T14:15:28.971Z] 14:15:28     INFO - GECKO(3067) | #01: mozilla::AudioCallbackDriver::SetStreamName(nsTSubstring<char> const&) [dom/media/GraphDriver.cpp:784]
[task 2023-10-27T14:15:28.971Z] 14:15:28     INFO - GECKO(3067) | #02: mozilla::MediaTrackGraphImpl::RunMessagesInQueue() [dom/media/MediaTrackGraph.cpp:1332]
[task 2023-10-27T14:15:28.971Z] 14:15:28     INFO - GECKO(3067) | #03: mozilla::MediaTrackGraphImpl::OneIterationImpl(long long, long long, mozilla::AudioMixer*) [dom/media/MediaTrackGraph.cpp:1588]
[task 2023-10-27T14:15:28.972Z] 14:15:28     INFO - GECKO(3067) | #04: mozilla::GraphRunner::Run() [dom/media/GraphRunner.cpp:139]
[task 2023-10-27T14:15:28.972Z] 14:15:28     INFO - GECKO(3067) | #05: nsThread::ProcessNextEvent(bool, bool*) [xpcom/threads/nsThread.cpp:1193]
[task 2023-10-27T14:15:28.972Z] 14:15:28     INFO - GECKO(3067) | #06: NS_ProcessNextEvent(nsIThread*, bool) [xpcom/threads/nsThreadUtils.cpp:480]
[task 2023-10-27T14:15:28.972Z] 14:15:28     INFO - GECKO(3067) | #07: mozilla::ipc::MessagePumpForNonMainThreads::Run(base::MessagePump::Delegate*) [ipc/glue/MessagePump.cpp:301]
[task 2023-10-27T14:15:28.973Z] 14:15:28     INFO - GECKO(3067) | #08: MessageLoop::Run() [ipc/chromium/src/base/message_loop.cc:346]
[task 2023-10-27T14:15:28.973Z] 14:15:28     INFO - GECKO(3067) | #09: nsThread::ThreadFunc(void*) [xpcom/threads/nsThread.cpp:372]
[task 2023-10-27T14:15:29.171Z] 14:15:29     INFO - GECKO(3067) | #10: _pt_root [nsprpub/pr/src/pthreads/ptthread.c:204]
[task 2023-10-27T14:15:29.172Z] 14:15:29     INFO - GECKO(3067) | #11: _pthread_start [/usr/lib/system/libsystem_pthread.dylib + 0x6109]
[task 2023-10-27T14:15:29.172Z] 14:15:29     INFO - GECKO(3067) | [Parent 3067, IPC I/O Parent] WARNING: Call to Send() failed: file /builds/worker/checkouts/gecko/ipc/glue/NodeChannel.cpp:198

Comment 1

[Iulian Moraru]

Hi Karl! Can you please take a look at this? Although this backfill range and retriggers don't show Bug 1860954 as the culprit, I think it might just be the cause for this and the 2 other assertion failure bugs that I've put in the "see also" section, but I would like to get your opinion on this.
Thank you!

Comment 2

[Intermittent Failures Robot]

15 failures in 3657 pushes (0.004 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 12
• mozilla-central: 1
• try: 2

Platform and build breakdown:

• macosx1015-64-qr: 6
  • debug: 6
• macosx1100-64-aarch64-qr: 7
  • debug: 7
• linux1804-64-qr: 2
  • debug: 2

For more details, see:
https://treeherder.mozilla.org/intermittent-failures/bugdetails?bug=1861709&startday=2023-10-23&endday=2023-10-29&tree=all

Comment 3

[Karl Tomlinson (:karlt)]

The AudioCallbackDriver transitions to ThreadRunning() while the fallback driver may still be in one of its iterations.

This is inconsistent with "If the fallback driver is active, this returns false" documented for AudioCallbackDriver::ThreadRunning().

Both are inconsistent with "GraphDriver's thread has started and the thread is running" documented for the overridden GraphDriver::ThreadRunning(). MediaTrackGraphImpl expects AudioCallbackDriver::ThreadRunning() to return true even when running from the fallback driver, though the difference in behavior is not detected there.

Comment 4

[Karl Tomlinson (:karlt)]

Attachment: Bug 1861709 permit AudioCallbackDriver::ThreadRunning() in FallbackWrapper::InDriverIteration() asserts r?pehrsons

AudioCallbackDriver will have a fallback driver thread running when in a
fallback driver iteration and sometimes also has an audio stream callback
thread running concurrently.

Comment 5

[Karl Tomlinson (:karlt)]

Attachment: Bug 1861709 simplify GraphDriver::InIteration() r?pehrsons

GraphInterface::InDriverIteration() now always checks aDriver->OnThread().
In the case of a fallback driver, this changes the check a little as
OnThread() is called on the AudioCallbackDriver instead of the fallback
driver, but I expect either IterationResult OneIteration() or its
IterationResult should be on the stack when asserting InIteration().

Depends on D192432

Comment 6

[Karl Tomlinson (:karlt)]

Attachment: Bug 1861709 replace AudioCallbackDriver::ThreadRunning() assertions that mean to test mAudioStreamState r?pehrsons

Depends on D192433

Comment 7

[Karl Tomlinson (:karlt)]

Attachment: Bug 1861709 align AudioCallbackDriver::ThreadRunning() behavior with its overridden GraphDriver method r?pehrsons

AudioCallbackDriver can have a thread running with a fallback driver or
an audio stream callback.

Asserts are using !ThreadRunning() and so are made stricter by this patch.

Depends on D192434

Comment 8

[Karl Tomlinson (:karlt)]

Thank you for the retriggers.

This showed up due to a new assertion introduced in https://hg.mozilla.org/mozilla-central/rev/899c4a0b9085b8781713f9fd29a5fdd91b55ade1#l1.275

Bug 1861805 and bug 1861495 are different to this one and I'm still investigating those.

Comment 9

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1860954

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/54251c8f8d09
permit AudioCallbackDriver::ThreadRunning() in FallbackWrapper::InDriverIteration() asserts r=pehrsons
https://hg.mozilla.org/integration/autoland/rev/0ea21c7c9ab7
replace AudioCallbackDriver::ThreadRunning() assertions that mean to test mAudioStreamState r=pehrsons

Comment 11

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/5215ac06e712
align AudioCallbackDriver::ThreadRunning() behavior with its overridden GraphDriver method r=pehrsons

Comment 12

[Natalia Csoregi [:nataliaCs]]

https://hg.mozilla.org/mozilla-central/rev/54251c8f8d09
https://hg.mozilla.org/mozilla-central/rev/0ea21c7c9ab7
https://hg.mozilla.org/mozilla-central/rev/5215ac06e712

Comment 14

[Intermittent Failures Robot]

36 failures in 3356 pushes (0.011 failures/push) were associated with this bug in the last 7 days.

** This failure happened more than 30 times this week! Resolving this bug is a high priority. **

** Try to resolve this bug as soon as possible. If unresolved for 2 weeks, the affected test(s) may be disabled. **

Repository breakdown:

• autoland: 28
• mozilla-central: 6
• try: 2

Platform and build breakdown:

• macosx1100-64-aarch64-qr: 13
  • debug: 13
• windows10-64-2009-qr: 1
  • debug: 1
• linux1804-64-qr: 2
  • debug: 2
• macosx1015-64-qr: 11
  • debug: 11
• android-em-7-0-x86_64-qr: 7
  • debug: 7
• windows11-64-2009-qr: 1
  • debug: 1
• windows11-32-2009-qr: 1
  • debug: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures/bugdetails?bug=1861709&startday=2023-10-30&endday=2023-11-05&tree=all