Closed Bug 1873388 Opened 5 months ago Closed 2 months ago

Crash in [@ <unknown in igd10um64gen11.dll> | CPixelShader::CLS::FinalConstruct]

Categories

(Core :: Graphics, defect)

Other
Windows
defect

Tracking

()

RESOLVED WONTFIX
Tracking Status
firefox123 --- affected

People

(Reporter: release-mgmt-account-bot, Unassigned)

References

(Blocks 1 open bug)

Details

(Keywords: crash)

Crash Data

Crash report: https://crash-stats.mozilla.org/report/index/6b55250f-e21a-4a56-b9d7-1b03b0231215

Reason: EXCEPTION_ACCESS_VIOLATION_READ

Top 10 frames of crashing thread:

0  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  
1  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  
2  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  
3  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  
4  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  
5  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  
6  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  
7  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  
8  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  
9  igd10um64gen11.DLL  <unknown in igd10um64gen11.dll>  

By querying Nightly crashes reported within the last 2 months, here are some insights about the signature:

  • First crash report: 2023-11-04
  • Process type: GPU
  • Is startup crash: No
  • Has user comments: No
  • Is null crash: Yes - 1 out of 4 crashes happened on null or near null memory address
Component: General → Graphics

Observations I thought notable:

  • Seems to be a startup crash; all reported crash instances are with 1 second or less of Uptime.
  • Windows' D3D11
  • Intel iGPU
  • All seem to crash in a background task for the compilation of PixelShaders in the D3D11 API being called by ANGLE, in a separate thread that I presume is internal to ANGLE.
  • Most crashes are from Firefox builds created in the last year. One of them extends back to Firefox 103.0, with a build ID dated to 2022-07-18. So, this doesn't seem like a new issue, or a widespread one.

NI'ing :jgilbert and :gw, who I know has familiarity with ANGLE bugs. I don't think that this is WebGL, but more likely WebRender's usage of ANGLE as a back end.

Severity: -- → S3
Flags: needinfo?(jgilbert)
Flags: needinfo?(gwatson)

The background thread is likely to be WR shaders, as we do async compilation on startup, and they are startup crashes.

One possibility is that the GPU driver (or ANGLE, I suppose) has a thread-safety bug, we could consider disabling background shader compilation on specific driver versions and seeing if that helps, though it would regress startup time so is a fairly heavy handed change.

Flags: needinfo?(gwatson)

WONTFIX unless volume increases.

Status: NEW → RESOLVED
Closed: 2 months ago
Flags: needinfo?(jgilbert)
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.