1709580 - Crash in [@ _tailMerge_d3dcompiler_47.dll]

Status: RESOLVED WORKSFORME

(Firefox Build System :: Toolchains, defect)

Description

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

_tailMerge_d3dcompiler_47.dll showed up in crash signatures starting with 90.0a1 20210429092605.

Signatures

Based on the the first builds affected, is this a regression from bug 1697215?

Crash report: https://crash-stats.mozilla.org/report/index/b3d5e2d7-59ba-4b32-901b-2f1c20210504

Reason: EXCEPTION_ILLEGAL_INSTRUCTION

Top 2 frames of crashing thread:

0  @0xabddd0e3d5 
1 xul.dll _tailMerge_d3dcompiler_47.dll 

Comment 1

[Andi [:andi]]

I'm confident it's only a coincidence.

Comment 2

[FredMcD]

https://support.mozilla.org/en-US/questions/1352606
bp-42b955a3-5ea9-4182-be86-fec190210930
Signature: RtlVirtualUnwind | _tailMerge_d3dcompiler_47.dll

Comment 3

[Sean Feng [:sefeng]]

I just added the crash signature to this bug, can you confirm it's still coincidence Andi?

Comment 4

[Gabriele Svelto [:gsvelto] (PTO)]

A quick note: IIRC _tailMerge_d3dcompiler_47.dll is the last public symbol that appears in the symbol file generated from XUL.dll. When you see it in a crash it means we've jumped into an area of memory that is still part of the mapping of XUL.dll but is outside of actual code. It's going to be hard to tell what's the problem and the signature might be hiding several different issues.

Comment 5

[Gabriele Svelto [:gsvelto] (PTO)]

I've opened a handful of crashes and they all appear to be non-actionable. The instruction pointer present in the context is bogus and the stack walker jumps into this area past our code in XUL.dll and stops there. This pointer - which is always in the second stack frame - gets the _tailMerge_d3dcompiler_47.dll symbol. I don't think there's anything actionable here and many, many of the crashes are coming from very old machines (10+ yo) so they're likely cause by bad hardware.

Comment 6

[Andi [:andi]]

(In reply to Sean Feng [:sefeng] from comment #3)

I just added the crash signature to this bug, can you confirm it's still coincidence Andi?

I concur with what gsvelto is saying.

Comment 7

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

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

• Top 10 content process crashes on beta

:glandium, could you consider increasing the severity of this top-crash bug?

For more information, please visit auto_nag documentation.

Comment 8

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

Based on the topcrash criteria, the crash signature linked to this bug is not a topcrash signature anymore.

For more information, please visit auto_nag documentation.

Comment 9

[Mike Hommey [:glandium]]

There doesn't seem to be any reports after 113.