1709026 - Crash in [@ __pthread_kill | abort | USC::CreateCompiler], likely Hackintosh related

Status: RESOLVED INACTIVE

(Core :: Graphics: WebRender, defect)

Description

[Gabriele Svelto [:gsvelto]]

Crash report: https://crash-stats.mozilla.org/report/index/69b921d1-4b02-4cf6-9ca2-3dc5c0210501

Reason: EXC_SOFTWARE / SIGABRT

Top 10 frames of crashing thread:

0 libsystem_kernel.dylib __pthread_kill 
1 libsystem_c.dylib abort 
2 libsystem_c.dylib __assert_rtn 
3 AppleIntelKBLGraphicsGLDriver USC::CreateCompiler 
4 AppleIntelKBLGraphicsGLDriver USC::CreateCompiler 
5 AppleIntelKBLGraphicsGLDriver GenContext::Initialize 
6 AppleIntelKBLGraphicsGLDriver glrInitializeContext 
7 libGPUSupportMercury.dylib gldCreateContext 
8 GLEngine gliCreateContextWithShared 
9 OpenGL CGLCreateContext 

This is a macOS-specific crash happening deep within Apple's driver code. Non-nightly versions of Firefox are affected so it doesn't appear to be a regression on our end (though it could be an old bug). It might also be a regression on Apple's side.

Comment 1

[Steven Michaud [:smichaud] (Retired)]

For what it's worth, a potentially interesting mac_crash_info section showed up in a recent crash report with this signature:

bp-d5062df2-9cf9-421f-9d71-1be0f0210713

    {
      "num_records": 1,
      "records": [
        {
          "message": "Assertion failed: (0), function CreateCompiler, file /Library/Caches/com.apple.xbs/Sources/GPUDriversIntel/GPUDriversIntel-14.7.8/GLRenderer/kbl/usc_interface.cpp, line 1676.\n",
          "module": "/usr/lib/system/libsystem_c.dylib"
        }
      ]
    }

Comment 2

[Lee Salzman [:lsalzman]]

After googling, https://github.com/alacritty/alacritty/issues/4375 seems to indicate this might be a problem with so-called "hackintosh" setups not recognizing the display hardware, and it seems like the user must properly configure their system to avoid it.

Comment 3

[Steven Michaud [:smichaud] (Retired)]

Interesting. There are quite a few hits for "usc_interface.cpp, line 1676", which is the last part of the "message" from the mac crash info from comment #1. These include the one you mentioned. All of those I looked at have to do with Hackintoshes.

Comment 4

[Steven Michaud [:smichaud] (Retired)]

And for what it's worth, all the crash reports associated with this bug have "usc_interface.cpp, line 1676" in their mac crash info:

https://crash-stats.mozilla.org/search/?proto_signature=~USC%3A%3ACreateCompiler&platform=Mac%20OS%20X&date=%3E%3D2021-06-04T03%3A32%3A00.000Z&date=%3C2021-12-04T03%3A32%3A00.000Z&_facets=signature&_facets=adapter_device_id&_facets=adapter_vendor_id&_facets=platform_version&_facets=mac_crash_info&_sort=-date&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-mac_crash_info

Edit: There are also some variants:
"usc_interface.cpp, line 1678"
"usc_interface.cpp, line 1626"

Comment 5

[Steven Michaud [:smichaud] (Retired)]

I haven't noticed many Mozilla bugs associated with weird hardware or settings. But there's at least one other proven one -- bug 1722827. And there may be more.

Comment 6

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

Closing because no crashes reported for 12 weeks.

Comment 7

[Steven Michaud [:smichaud] (Retired)]

These are still happening.

Comment 8

[Glenn Watson [:gw]]

Given the volume and lack of repro steps, there is nothing we can do about this. Closing for now, but please reopen if volume or relevant information changes.