1783212 - macOS Crash in [@ pthread_kill | abort | gpusKillClientExt]

Status: NEW

(Core :: Graphics, defect)

Description

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

101 crashes for various Firefox versions (mostly 103.0, 103.0.1 and 91.12.0esr), all on macOS. The platform version of one report I checked had been released before 2022-08-02. New graphics driver?

Crash report: https://crash-stats.mozilla.org/report/index/365ad625-53b8-4a27-98d0-f60a40220802

Reason: EXC_SOFTWARE / SIGABRT

Top 10 frames of crashing thread:

0 libsystem_kernel.dylib __pthread_kill 
1 libsystem_pthread.dylib pthread_kill 
2 libsystem_c.dylib abort 
3 libGPUSupportMercury.dylib gpusGenerateCrashLog.cold.1 
4 libGPUSupportMercury.dylib gpusGenerateCrashLog 
5 AppleIntelKBLGraphicsGLDriver gpusKillClientExt 
6 libGPUSupportMercury.dylib gpusSubmitDataBuffers 
7 AppleIntelKBLGraphicsGLDriver IntelCommandBuffer::getNew 
8 AppleIntelKBLGraphicsGLDriver intelSubmitCommands 
9 AppleIntelKBLGraphicsGLDriver gldClearFramebufferData 

Comment 1

[Steven Michaud [:smichaud] (Retired)]

These aren't new. Most of them have already been reported under various signatures including gpusGenerateCrashLog.cold.1 or gpusGenerateCrashLog.cold.

https://crash-stats.mozilla.org/search/?proto_signature=~gpusKillClientExt&platform=Mac%20OS%20X&date=%3E%3D2022-07-28T14%3A59%3A00.000Z&date=%3C2022-08-04T14%3A59%3A00.000Z&_facets=signature&_facets=platform_version&_facets=proto_signature&_facets=mac_crash_info&_sort=-date&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-signature

Comment 2

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

The bug is linked to a topcrash signature, which matches the following criterion:

• Top 5 desktop browser crashes on Mac on release

For more information, please visit auto_nag documentation.

Comment 3

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

The bug is linked to a topcrash signature, which matches the following criterion:

• Top 5 desktop browser crashes on Mac on release (startup)

For more information, please visit auto_nag documentation.

Comment 4

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

Based on the topcrash criteria, the crash signatures linked to this bug are not in the topcrash signatures anymore.

For more information, please visit auto_nag documentation.

Comment 5

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

Sorry for removing the keyword earlier but there is a recent change in the ranking, so the bug is again linked to a topcrash signature, which matches the following criteria:

• Top 20 desktop browser crashes on release (startup)
• Top 5 desktop browser crashes on Mac on release (startup)

For more information, please visit BugBot documentation.

Comment 6

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

Based on the topcrash criteria, the crash signatures linked to this bug are not in the topcrash signatures anymore.

For more information, please visit BugBot documentation.

Comment 8

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

Copying crash signatures from duplicate bugs.

Comment 9

[Ashley Hale [:ahale]]

During triage this came up in top crashes, we plan to a second look in next triage meeting, so adding to gfx-triage this time.

Among the many crashes that I looked at, I noticed that a couple crashes have excessive surface sizes in WebRender:

• https://crash-stats.mozilla.org/report/index/ab98e53f-3a4f-4fb6-92bc-1805e0250425 - GfxCriticalError says: |[0][GFX1-]: Failed to allocate a surface due to invalid size (CDT) Size(0,0) (t=1.22645e+06) |[1][GFX1-]: Failed to allocate a surface due to invalid size (CDT) Size(0,0) (t=1.22647e+06)
• https://crash-stats.mozilla.org/report/index/c9b278a3-83e3-4067-8ba4-d4c9e0250422 - GfxCriticalError says: |[0][GFX1-]: Failed to allocate a surface due to invalid size (CDT) Size(0,0) (t=46875.5)

I suspect there are more themes we can look at here. A substantial portion of the crashes are WebGL when recreating the context, and a similarly large portion are webRender when recreating the context - it would be nice if we knew what happened before the context was lost though.

Comment 10

[Steven Michaud [:smichaud] (Retired)]

(In reply to Ashley Hale [:ahale] from comment #9)

During triage this came up in top crashes, we plan to a second look in next triage meeting, so adding to gfx-triage this time.

Among the many crashes that I looked at, I noticed that a couple crashes have excessive surface sizes in WebRender:

• https://crash-stats.mozilla.org/report/index/ab98e53f-3a4f-4fb6-92bc-1805e0250425 - GfxCriticalError says: |[0][GFX1-]: Failed to allocate a surface due to invalid size (CDT) Size(0,0) (t=1.22645e+06) |[1][GFX1-]: Failed to allocate a surface due to invalid size (CDT) Size(0,0) (t=1.22647e+06)
• https://crash-stats.mozilla.org/report/index/c9b278a3-83e3-4067-8ba4-d4c9e0250422 - GfxCriticalError says: |[0][GFX1-]: Failed to allocate a surface due to invalid size (CDT) Size(0,0) (t=46875.5)

I suspect there are more themes we can look at here. A substantial portion of the crashes are WebGL when recreating the context, and a similarly large portion are webRender when recreating the context - it would be nice if we knew what happened before the context was lost though.

Most (almost all) of these "graphics critical errors" are on very old versions of macOS. But a few do exist on more recent versions:

https://crash-stats.mozilla.org/search/?proto_signature=~gpusKillClientExt&graphics_critical_error=~Failed%20to%20allocate%20a%20surface%20due%20to%20invalid%20size&platform=Mac%20OS%20X&date=%3E%3D2024-10-25T04%3A26%3A00.000Z&date=%3C2025-04-25T04%3A26%3A00.000Z&_facets=signature&_facets=platform_version&_facets=graphics_critical_error&_sort=-date&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-platform_version

Comment 11

[Steven Michaud [:smichaud] (Retired)]

Here's a better (wider) search for "graphics critical errors" associated with this bug's crashes. I've eliminated crashes on macOS 10.15 and earlier, since those crashes are unlikely to concern Mozilla now.

https://crash-stats.mozilla.org/search/?proto_signature=~gpusKillClientExt&platform_version=%21%5E10.&graphics_critical_error=%21__null__&platform=Mac%20OS%20X&date=%3E%3D2024-10-25T17%3A29%3A00.000Z&date=%3C2025-04-25T17%3A29%3A00.000Z&_facets=signature&_facets=platform_version&_sort=-date&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform_version&_columns=graphics_critical_error#crash-reports

Comment 12

[Ashley Hale [:ahale]]

GPU process will at least change this situation and hopefully improve it, so taking out of triage

Comment 13

[Steven Michaud [:smichaud] (Retired)]

This bug should have been fixed by bug 1889757, but it doesn't seem that it was. See bug 1838649 comment #16.

Comment 14

[Brad Werth [:bradwerth]]

(In reply to Steven Michaud [:smichaud] (Retired) from comment #13)

This bug should have been fixed by bug 1889757, but it doesn't seem that it was. See bug 1838649 comment #16.

GPU process is not on by default, and will not be until Bug 1985082 lands. That Bug will turn it on for Nightly, and then another Bug (not yet filed) will turn it on for all channels.

Comment 15

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

Based on the topcrash criteria, the crash signatures linked to this bug are not in the topcrash signatures anymore.

For more information, please visit BugBot documentation.