Closed Bug 1737365 Opened 3 years ago Closed 2 years ago

Crash in [@ mozilla::layers::SurfacePoolCA::LockedPool::ReturnSurfaceToPool]

Categories

(Core :: Graphics: Layers, defect)

x86_64
macOS
defect

Tracking

()

RESOLVED WORKSFORME
Tracking Status
firefox-esr78 --- unaffected
firefox-esr91 --- unaffected
firefox93 --- unaffected
firefox94 --- wontfix
firefox95 --- wontfix

People

(Reporter: mccr8, Unassigned)

References

Details

(Keywords: crash, regression, Whiteboard: [Bug 1736446 is possible fix?])

Crash Data

Crash report: https://crash-stats.mozilla.org/report/index/0c2bf6e2-5e13-4354-8e19-769f80211022

MOZ_CRASH Reason: MOZ_RELEASE_ASSERT(inUseEntryIter != mInUseEntries.end())

Top 10 frames of crashing thread:

0 XUL mozilla::layers::SurfacePoolCA::LockedPool::ReturnSurfaceToPool gfx/layers/SurfacePoolCA.mm:181
1 XUL mozilla::layers::NativeLayerCA::~NativeLayerCA gfx/layers/NativeLayerCA.mm:672
2 XUL mozilla::layers::NativeLayerCA::~NativeLayerCA gfx/layers/NativeLayerCA.mm:662
3 XUL mozilla::layers::NativeLayerRootCA::SetWindowIsFullscreen gfx/layers/NativeLayerCA.mm:491
4 XUL nsCocoaWindow::DoMakeFullScreen widget/cocoa/nsCocoaWindow.mm:1675
5 XUL nsGlobalWindowOuter::SetWidgetFullscreen dom/base/nsGlobalWindowOuter.cpp:4649
6 XUL FullscreenTransitionTask::Run dom/base/nsGlobalWindowOuter.cpp:4406
7 XUL mozilla::TaskController::DoExecuteNextTaskOnlyMainThreadInternal xpcom/threads/TaskController.cpp:770
8 XUL NS_ProcessPendingEvents xpcom/threads/nsThreadUtils.cpp:432
9 XUL nsAppShell::ProcessGeckoEvents widget/cocoa/nsAppShell.mm:505

While I was helping to investigate bug 1734854 by doing a bisection, I had the space station stream open from Bug 1734854 comment 13, and occasionally popped it in or out of full screen (to try to make sure I remembered what the behavior was like in my original profile). A few minutes after I finished doing the bisection, I went back to my original browser and was again entering or exiting full screen on the tab for that video, and I hit this release assert.

I'm on a 10-22 Nightly build.

OS: Unspecified → macOS
Hardware: Unspecified → x86_64

It looks like this release assert is also being hit on beta, under a slightly different signature: bp-fe292055-9bad-41b4-86e6-3380b0211021

The crashes all have DoMakeFullScreen in the stacks.

Crash Signature: [@ mozilla::layers::SurfacePoolCA::LockedPool::ReturnSurfaceToPool] → [@ mozilla::layers::SurfacePoolCA::LockedPool::ReturnSurfaceToPool] [@ mozilla::layers::SurfacePoolCA::ReturnSurfaceToPool ]

If this is only occurring with the combination of macOS video fullscreen shortly after a fullscreen transition, then the patch for Bug 1736446 may be relevant. Can anyone consistently reproduce in a local build?

Unfortunately I haven't been able to reproduce the crash.

For what it is worth, the earlier build I found with this release assert with DoMakeFullScreen in the stack was this one (it shows up on much older branches, but the stacks look different) from the 20210929155324 build: bp-71a2bc5b-539b-4349-b855-335200210930

Keywords: regression
Whiteboard: [Bug 1736446 is possible fix?]

Low-volume crash so far on Beta94, but holding off on a full wontfix until we see what it does on Release next week.

This hasn't shown up since 94.

Status: NEW → RESOLVED
Closed: 2 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.