1459233 - Crash in mozilla::layers::SyncObjectD3D11Host::Synchronize

Status: RESOLVED FIXED

(Core :: Graphics: WebRender, defect, P2)

Description

[Marcia Knous [:marcia]]

This bug was filed from the Socorro interface and is
report bp-a5672ed8-ee14-450f-a100-dd2e50180423.
=============================================================

Seen while looking at Windows nightly crash data - it looks as if crashes started when nightly was still in 60: https://bit.ly/2HW3ZIU. In the last 7 days, 12 crashes/8 installs.

Moz Crash Reason: MOZ_CRASH(GFX: D3D11 normal status timeout)

I put it in this component since it looks as if code was touched in a few bugs such as Bug 1357299 - feel free to move it if that is not correct.

Top 10 frames of crashing thread:

0 xul.dll mozilla::layers::SyncObjectD3D11Host::Synchronize gfx/layers/d3d11/TextureD3D11.cpp:1710
1 xul.dll mozilla::layers::MLGDeviceD3D11::Synchronize gfx/layers/d3d11/MLGDeviceD3D11.cpp:1834
2 xul.dll mozilla::layers::LayerManagerMLGPU::RenderLayers gfx/layers/mlgpu/LayerManagerMLGPU.cpp:367
3 xul.dll mozilla::layers::LayerManagerMLGPU::Composite gfx/layers/mlgpu/LayerManagerMLGPU.cpp:314
4 xul.dll mozilla::layers::LayerManagerMLGPU::EndTransaction gfx/layers/mlgpu/LayerManagerMLGPU.cpp:273
5 xul.dll mozilla::layers::CompositorBridgeParent::CompositeToTarget gfx/layers/ipc/CompositorBridgeParent.cpp:1022
6 xul.dll mozilla::layers::CompositorVsyncScheduler::Composite gfx/layers/ipc/CompositorVsyncScheduler.cpp:243
7 xul.dll mozilla::detail::RunnableMethodImpl<mozilla::layers::CompositorVsyncScheduler*, void  xpcom/threads/nsThreadUtils.h:1215
8 xul.dll MessageLoop::DoWork ipc/chromium/src/base/message_loop.cc:535
9 xul.dll base::MessagePumpForUI::DoRunLoop ipc/chromium/src/base/message_pump_win.cc:210

=============================================================

Comment 1

[Sotaro Ikeda [:sotaro]]

The crash stat in comment 0 has LayerManagerMLGPU, it means WebRender is not used.

Comment 2

[Darkspirit]

WebRender was affected by this three times within the last month:
https://crash-stats.mozilla.com/search/?signature=~mozilla%3A%3Alayers%3A%3ASyncObjectD3D11Host%3A%3ASynchronize&app_notes=~WR%2B&product=Firefox&date=%3E%3D2018-04-09T12%3A31%3A39.000Z&date=%3C2018-05-09T12%3A31%3A39.000Z&_sort=-date&_facets=signature&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#crash-reports

Comment 3

[Sotaro Ikeda [:sotaro]]

Thanks for checking!

Comment 4

[Sotaro Ikeda [:sotaro]]

Problem with WebRender could be mitigated by Bug 1364504.

Comment 5

[Sotaro Ikeda [:sotaro]]

Attachment: patch - Make SyncObjectD3D11Host::Synchronize() fallible with WebRender

Comment 6

[Sotaro Ikeda [:sotaro]]

attachment 9013534 [details] [diff] [review] makes SyncObjectD3D11Host::Synchronize() fallible like SyncObjectD3D11Client::Synchronize(bool aFallible).

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/4505ab264626
Make SyncObjectD3D11Host::Synchronize() fallible with WebRender r=mattwoodrow

Comment 8

[Andreea Pavel [:apavel]]

https://hg.mozilla.org/mozilla-central/rev/4505ab264626

Comment 9

[Tiberius Oros[:tiberius_oros]]

https://hg.mozilla.org/mozilla-central/rev/4505ab264626

Comment 10

[Andrei Ciure[:aciure]]

There is a perma fail for this failure:
https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&resultStatus=testfailed%2Cbusted%2Cexception&classifiedState=unclassified&fromchange=f9572cfd7bac75d4493f44206cd27958d74269d0&selectedJob=203012642&searchStr=windows%2C10%2Cx64%2Cquantumrender%2Copt%2Creftests%2Cwith%2Ce10s%2Ctest-windows10-64-qr%2Fopt-reftest-e10s-1%2Cr-e10s%28r1%29

:sotaro can you take a look at this?
We will back it out in the next merge if it's not fixed.

Comment 11

[Sotaro Ikeda [:sotaro]]

I looked into the code, but it is not clear whey it causes a perma fail. It is OK to backout.

Comment 12

[Sotaro Ikeda [:sotaro]]

Bug 1489327 already exists. Not sure why it becomes a perma fail.

Comment 13

[Sotaro Ikeda [:sotaro]]

If the change causes the perma fail, it means that GPU process crashed because of D3D11 timeout on the D3D11 sync lock.

Comment 14

[Pulsebot]

Backout by [e-mail]:
https://hg.mozilla.org/mozilla-central/rev/775ad70be301
Backed out 1 changesets for continuation-opacity.html failures on Win 10 qr a=backout

Comment 15

[Andrei Ciure[:aciure]]

Backed out 1 changesets (bug 1459233) for continuation-opacity.html failures on Win 10 qr

push that caused the backout: https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&searchStr=94c3621993359b84f08a9a98ee45d74673a841a7&selectedJob=203033297&revision=4505ab264626d70bdc34ed6e6f9d7b13ae074b48

failures: https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&searchStr=94c3621993359b84f08a9a98ee45d74673a841a7&selectedJob=203033297&revision=fcfe79760aa85edf977bb3753321c2858347505f

backout: https://hg.mozilla.org/mozilla-central/rev/775ad70be30122138b07b3928e1421187046afb0

Comment 16

[Sotaro Ikeda [:sotaro]]

(In reply to Sotaro Ikeda [:sotaro] from comment #13)
> If the change causes the perma fail, it means that GPU process crashed
> because of D3D11 timeout on the D3D11 sync lock.

Hmm, it seems not called during the test on try. Then the change seems not actually related.

Comment 17

[Sotaro Ikeda [:sotaro]]

Attachment: patch - Make SyncObjectD3D11Host::Synchronize() fallible with WebRender

Rebased.

Comment 18

[Sotaro Ikeda [:sotaro]]

unexpected fail of continuation-opacity.html was addressed by Bug 1496324 

https://treeherder.mozilla.org/#/jobs?repo=try&revision=60b23def22cb1f9f8e04387ddd9e652f2a35064d

Comment 19

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/2558c8ea6ad2
Make SyncObjectD3D11Host::Synchronize() fallible with WebRender r=mattwoodrow

Comment 20

[Cosmin Sabou [:CosminS]]

https://hg.mozilla.org/mozilla-central/rev/2558c8ea6ad2