1906490 - Crash in [@ <unknown in amdxx64.dll> | CVideoProcessor::Blt]

Status: NEW

(Core :: Graphics: WebRender, defect)

Description

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

Crash report: https://crash-stats.mozilla.org/report/index/22d81949-9c66-4a1c-a2fc-0a0ce0240614

Reason: EXCEPTION_ACCESS_VIOLATION_READ

Top 10 frames of crashing thread:

0  amdxx64.dll  <unknown in amdxx64.dll>  
1  amdxx64.dll  <unknown in amdxx64.dll>  
2  amdxx64.dll  <unknown in amdxx64.dll>  
3  amdxx64.dll  <unknown in amdxx64.dll>  
4  amdxx64.dll  <unknown in amdxx64.dll>  
5  amdxx64.dll  <unknown in amdxx64.dll>  
6  amdxx64.dll  <unknown in amdxx64.dll>  
7  amdxx64.dll  <unknown in amdxx64.dll>  
8  amdxx64.dll  <unknown in amdxx64.dll>  
9  amdxx64.dll  <unknown in amdxx64.dll>  

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

• First crash report: 2024-05-20
• Process type: GPU
• Is startup crash: No
• Has user comments: No
• Is null crash: Yes - 6 out of 7 crashes happened on null or near null memory address

Comment 1

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

The Bugbug bot thinks this bug should belong to the 'Core::Graphics' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Comment 2

[Timothy Nikkel (:tnikkel)]

8 total crashes since march.

Comment 3

[Jamie Nicol [:jnicol]]

It is slightly curious that all the crashes are on Nightly, though multiple versions. The sample size may be too low to draw any conclusions from that, and most the crashes are 129. Sotaro, have there been any recent changes to our DirectComposition usage that are either nightly only for the past few releases, or a new change in 129?

Comment 4

[Sotaro Ikeda [:sotaro]]

I wonder if Bug 1899450 might be related to the problem.

Comment 5

[Sotaro Ikeda [:sotaro]]

GraphicsCriticalError of crash reports often has the following error. I wondered if D3D11Device may not have been working when the problem happened.

|[G0][GFX1-]: RenderDXGITextureHost init stream failed

https://searchfox.org/mozilla-central/rev/f7b41fc41c5505db507d076c16961a8da67dd318/gfx/webrender_bindings/RenderD3D11TextureHost.cpp#349

Comment 6

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

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

• Top 5 GPU process crashes on release

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

For more information, please visit BugBot documentation.

Comment 7

[Gabriele Svelto [:gsvelto]]

The sudden crash spike here is an effect of bug 1863288. We're now properly scraping the drivers debug information so we have coherent stacks, which in turn are coalescing under this signature. The volume here was already present, but spread through other signatures which we were not tracking (and which were not too useful), it's not an increase per se.

Comment 8

[Glenn Watson [:gw] | PTO until 29th April]

Unlikely to be anything actionable without a local repro.

Comment 9

[Gabriele Svelto [:gsvelto]]

This is spiking a bit in nightly and there's a few things that could make it easy to repro locally, only three drivers are affected:

• 32.0.12033.1030
• 32.0.12019.1028
• 32.0.12011.1036

The bulk of the crashes are happening with the Navi 31 and 32 families of GPUs (Radeon RX 7700/7800/7900) and the vast majority of the crashes is happening while watching videos on https://9gag.com and https://9gag.com/trending

Given that this appears to be a driver bug, do we have a contact with AMD to inform them?

Comment 10

[Gian-Carlo Pascutto [:gcp]]

Bob, see above, can we point AMD at this?

Comment 11

[Sotaro Ikeda [:sotaro]]

Crash happened also on release. Then the crash is not limited to zero video copy with AMD GPUs.

Comment 12

[Bob Hood [:bhood]]

I'll loop in Paul Blinzer (AMD) here just for an FYI, but from Sotaro's comment, it seems like this isn't isolated to AMD GPUs. The crash volume appears to currently be low, so I'll let this one simmer until the WR team has more cycles to make progress.

Comment 13

[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 BugBot documentation.

Comment 14

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

Redirect a needinfo that is pending on an inactive user to the triage owner.
:gw, since the bug has recent activity, could you have a look please?

For more information, please visit BugBot documentation.