1729648 - Crash in [@ __pthread_kill | pthread_kill | abort | MTLReportFailure.cold.1]

Status: RESOLVED WORKSFORME

(Core :: Graphics: WebGPU, defect)

Description

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

15 crashes on 5+ macOS 11 installations, oldest build is Firefox 93.0a1 20210904094024.

Submitted URLs indicate an issue related to WebGPU.

Maybe Fission related. (DOMFissionEnabled=1)

Crash report: https://crash-stats.mozilla.org/report/index/33412143-79e6-46e0-8b65-3a3c60210908

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 libsystem_c.dylib __assert_rtn 
4 Metal MTLReportFailure.cold.1 
5 Metal MTLReportFailure 
6 Metal _MTLMessageContextEnd 
7 Metal -[MTLTextureDescriptorInternal validateWithDevice:] 
8 AppleIntelICLGraphicsMTLDriver validateTextureDescriptor 
9 AppleIntelICLGraphicsMTLDriver -[MTLIGAccelDevice newTextureWithDescriptor:] 

Comment 1

[Calixte Denizet (:calixte)]

There are 20 crashes (from 7 installations) in nightly 93 starting with buildid 20210904094024. In analyzing the backtrace, the regression may have been introduced by patch [1] to fix bug 1726626.

[1] https://hg.mozilla.org/mozilla-central/rev?node=ef5dc3e04e5f

Comment 2

[Steven Michaud [:smichaud] (Retired)]

All of these crashes (so far) have one of the following in "mac crash info":

    {
      "num_records": 1,
      "records": [
        {
          "message": "-[_MTLCommandEncoder dealloc]:131: failed assertion `Command encoder released without endEncoding'\n",
          "module": "/usr/lib/system/libsystem_c.dylib"
        }
      ]
    }

    {
      "num_records": 1,
      "records": [
        {
          "message": "-[MTLTextureDescriptorInternal validateWithDevice:]:1248: failed assertion `Texture Descriptor Validation\nMTLTextureDescriptor has width of zero.\nMTLTextureDescriptor has height of zero.\n'\n",
          "module": "/usr/lib/system/libsystem_c.dylib"
        }
      ]
    }

Comment 3

[Dzmitry Malyshau [:kvark]]

Changing severity to S2 because of it's nightly only and not enabled by default.
So far, I found the reason of the first crash "Command encoder released without endEncoding" and have a fix coming, but not the second yet.

Comment 4

[Dzmitry Malyshau [:kvark]]

First half is addressed by https://phabricator.services.mozilla.com/D125103

Comment 5

[Steven Michaud [:smichaud] (Retired)]

D125103 seems to have landed, but there are still lots of crashes with the "Command encoder released without endEncoding" error:

https://crash-stats.mozilla.org/search/?mac_crash_info=~Command%20encoder%20released%20without%20endEncoding&version=95.0a1&platform=Mac%20OS%20X&date=%3E%3D2021-10-08T17%3A26%3A00.000Z&date=%3C2021-10-15T17%3A26%3A00.000Z&_facets=signature&_sort=-date&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-signature

Comment 6

[Dzmitry Malyshau [:kvark]]

Fixed by https://phabricator.services.mozilla.com/D132058

Comment 7

[Steven Michaud [:smichaud] (Retired)]

These crashes are still happening, though only on mozilla-central nightlies:

https://crash-stats.mozilla.org/search/?mac_crash_info=~Command%20encoder%20released%20without%20endEncoding&platform=Mac%20OS%20X&date=%3E%3D2021-12-18T03%3A50%3A00.000Z&date=%3C2022-03-18T03%3A50%3A00.000Z&_facets=signature&_facets=version&_sort=-date&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-version

Edit: They do seem to be happening at a lower volume, though.

Comment 8

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

The bug assignee is inactive on Bugzilla, so the assignee is being reset.

Comment 9

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

Closing because no crashes reported for 12 weeks.