1547645 - Crash in [@ @0x0 | nvwgf2umx_cfg.dll | CSampler::GetDesc]

Status: RESOLVED WORKSFORME

(Core :: Graphics, defect, P5)

Description

[Gabriele Svelto [:gsvelto] (PTO)]

This bug is for crash report bp-ac622f0b-139d-4409-9acb-15e130190428.

Top 10 frames of crashing thread:

0  @0x0 
1 nvwgf2umx_cfg.dll nvwgf2umx_cfg.dll@0xf25073 
2 d3d11.dll CSampler::GetDesc 
3 nvwgf2umx_cfg.dll nvwgf2umx_cfg.dll@0xba4082 
4 mozglue.dll void* arena_t::MallocSmall memory/build/mozjemalloc.cpp:2773
5 mozglue.dll void* BaseAllocator::malloc memory/build/mozjemalloc.cpp:3896
6  @0xb4f7bff027 
7 mozglue.dll void* BaseAllocator::malloc memory/build/mozjemalloc.cpp:3896
8 mozglue.dll static void* Allocator<MozJemallocBase>::malloc memory/build/malloc_decls.h:38
9 mozglue.dll moz_xmalloc memory/mozalloc/mozalloc.cpp:68

This is a new crash happening in the Renderer thread of the GPU process. All the reports are using either nVidia GeForce GTX 1070 or 1080 cards with driver version 25.21.14.2531 so it looks very specific to this combination.

The crashes are happening only on Nightly starting with buildid 20190417093858.

The backtrace is unfortunately almost useless because the unwinding machinery is using stack-scanning to crawl out of the vendor libraries and it's not being very successful at that.

Comment 1

[Timothy Nikkel (:tnikkel)]

Any chance we could get a pushlog with the changes between the nightlies where this started?

Comment 2

[Gabriele Svelto [:gsvelto] (PTO)]

I've double-checked the reports again before looking for the pushlog and there's something I hadn't noticed: while install times are different these all seem to come from a single machine. I was mislead by the different installation times into thinking that this was a valid issue but it's just a single-installation crash from a faulty machine.

Comment 3

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

Closing because no crashes reported for 12 weeks.