1521774 - Crash in mozilla::gfx::Factory::GetD2DDeviceContext

Status: RESOLVED FIXED

(Core :: Graphics, defect, P1)

Description

[Marcia Knous [:marcia]]

This bug is for crash report bp-b6685b06-5afa-4934-a3f1-4e34a0190121.

Small volume Win regression which started in 20190119095933: https://bit.ly/2HsTD7i.

Code was touched in Bug 1519760. ni on :bas

Top 10 frames of crashing thread:

0 xul.dll mozilla::gfx::Factory::GetD2DDeviceContext gfx/2d/Factory.cpp:875
1 xul.dll mozilla::gfx::DrawTargetD2D1::OptimizeSourceSurface gfx/2d/DrawTargetD2D1.cpp:2157
2 xul.dll static struct nsLayoutUtils::SurfaceFromElementResult nsLayoutUtils::SurfaceFromElement layout/base/nsLayoutUtils.cpp:7277
3 xul.dll nsLayoutUtils::SurfaceFromElement layout/base/nsLayoutUtils.cpp:7424
4 xul.dll mozilla::dom::CanvasRenderingContext2D::DrawImage dom/canvas/CanvasRenderingContext2D.cpp:4658
5 xul.dll static bool mozilla::dom::CanvasRenderingContext2D_Binding::drawImage dom/bindings/CanvasRenderingContext2DBinding.cpp:2621
6 xul.dll mozilla::dom::binding_detail::GenericMethod<mozilla::dom::binding_detail::NormalThisPolicy, mozilla::dom::binding_detail::ThrowExceptions> dom/bindings/BindingUtils.cpp:3124
7 xul.dll js::InternalCallOrConstruct js/src/vm/Interpreter.cpp:533
8 xul.dll static bool InternalCall js/src/vm/Interpreter.cpp:588
9 xul.dll static bool Interpret js/src/vm/Interpreter.cpp:3068

Comment 2

[Marcia Knous [:marcia]]

Marking 67 as affected. This is visible in the 66 betas as well, in small volume.

Some URLs:

• https://www.forbes.com/sites/jasonevangelho/2018/12/10/meet-the-linux-desktop-that-blows-away-windows-10-and-macos/#4c4363a85000
• https://vk.com/feed
• https://www.youtube.com/watch?v=kuoN1izAjSc

Correlations for Firefox Nightly:
(35.71% in signature vs 99.95% overall) ipc_message_name = null
(42.86% in signature vs 00.67% overall) GFX_ERROR "Compositors might be mixed (" = true [120.00% vs 00.55% if startup_crash = null]
(100.0% in signature vs 26.23% overall) address = 0x0
(100.0% in signature vs 14.36% overall) GFX_ERROR ": " = true [100.0% vs 20.27% if adapter_vendor_id = 0x8086]
(100.0% in signature vs 33.70% overall) reason = EXCEPTION_ACCESS_VIOLATION_READ
(71.43% in signature vs 02.41% overall) GFX_ERROR "Receive IPC close with reason=AbnormalShutdown" = true [85.71% vs 03.81% if adapter_vendor_id = 0x10de]
(35.71% in signature vs 00.27% overall) Module "nvapi64.dll" = true

Comment 3

[Liz Henry (:lizzard) (relman/hg->git project)]

Maybe related to bug 1524554 and its friend bug 1521368. If so, we should see a significant drop in the beta 5 crash rate.

Comment 4

[Bas Schouten (:bas.schouten)]

I don't see this particular crash in the top50 for Beta 5.

Comment 5

[[:philipp]]

it's the #17 content crash in 66.0b5 (0.6% of all tab crashes)

Comment 6

[Bas Schouten (:bas.schouten)]

Yep, I see it's appeared in the top 50 now. The fix for this is pretty easy so I'll put it up here, and we can decide if it needs uplifting.

Comment 7

[Bas Schouten (:bas.schouten)]

Attachment: Bug 1521774: Ensure a DC is available when using the generic dc on the main thread. r=rhunt

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/ab254a4faafe
Ensure a DC is available when using the generic dc on the main thread. r=rhunt

Comment 9

[Noemi Erli[:noemi_erli]]

https://hg.mozilla.org/mozilla-central/rev/ab254a4faafe

Comment 10

[Liz Henry (:lizzard) (relman/hg->git project)]

Bas, can you request beta uplift? Thanks!

Comment 11

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

Changing the priority to p1 as the bug is tracked by a release manager for the current beta.
See How Do You Triage for more information

Comment 12

[Bas Schouten (:bas.schouten)]

Comment on attachment 9042462 [details]
Bug 1521774: Ensure a DC is available when using the generic dc on the main thread. r=rhunt

Beta/Release Uplift Approval Request

Feature/Bug causing the regression

Bug 1519760

User impact if declined

Content process crash on some decide resets

Is this code covered by automated tests?

No

Has the fix been verified in Nightly?

Yes

Needs manual test from QE?

No

If yes, steps to reproduce

List of other uplifts needed

None

Risk to taking this patch

Low

Why is the change risky/not risky? (and alternatives if risky)

Simple null pointer check

String changes made/needed

None

Comment 13

[Liz Henry (:lizzard) (relman/hg->git project)]

Comment on attachment 9042462 [details]
Bug 1521774: Ensure a DC is available when using the generic dc on the main thread. r=rhunt

Top crash fix, let's uplift for beta 8.

Comment 14

[Daniel Varga [:dvarga]]

https://hg.mozilla.org/releases/mozilla-beta/rev/1bbd5e11e62b

Comment 15

[[:philipp]]

the number of submitted crash reports on beta hasn't significantly improved after the patch has landed unfortunately.

Comment 16

[Liz Henry (:lizzard) (relman/hg->git project)]

Is bug 1526045 related?

Comment 17

[Bas Schouten (:bas.schouten)]

Attachment: Bug 1521774 - Part 2: Ensure a device is available when attempting to create a device context. r=rhunt

Comment 18

[Liz Henry (:lizzard) (relman/hg->git project)]

Is this ready to land?

Comment 19

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/0d10f1a69e9a
Part 2: Ensure a device is available when attempting to create a device context. r=rhunt

Comment 20

[Natalia Csoregi [:nataliaCs]]

https://hg.mozilla.org/mozilla-central/rev/0d10f1a69e9a

Comment 21

[Bas Schouten (:bas.schouten)]

Comment on attachment 9048115 [details]
Bug 1521774 - Part 2: Ensure a device is available when attempting to create a device context. r=rhunt

Beta/Release Uplift Approval Request

• Feature/Bug causing the regression: Bug 1519760
• User impact if declined: Crashes
• Is this code covered by automated tests?: Yes
• Has the fix been verified in Nightly?: Yes
• Needs manual test from QE?: No
• If yes, steps to reproduce:
• List of other uplifts needed: None
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): Simple null check
• String changes made/needed: None

Comment 22

[Liz Henry (:lizzard) (relman/hg->git project)]

Comment on attachment 9048115 [details]
Bug 1521774 - Part 2: Ensure a device is available when attempting to create a device context. r=rhunt

Plugging away at these D2D crashes. OK to uplift.
We'll see how it does in the RC build next week.

Comment 23

[Narcis Beleuzu [:NarcisB]]

https://hg.mozilla.org/releases/mozilla-beta/rev/067b9599e2aa