1718438 - Intermittent Windows 10 AArch64 qr opt [tier 2] bugs/76331-1.html == bugs/76331-1-ref.html | image comparison, max difference: 255, number of differing pixels: 7043 ("RenderThread detected a device reset in PostUpdate" "Device reset due to WR device")

Status: RESOLVED WORKSFORME

(Core :: Graphics: WebRender, defect, P5)

Description

[Treeherder Bug Filer]

Filed by: imoraru [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=343971333&repo=mozilla-central
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/bsMZKyVeQDW4OXTWtgjfkA/runs/0/artifacts/public/logs/live_backing.log
Reftest URL: https://hg.mozilla.org/mozilla-central/raw-file/tip/layout/tools/reftest/reftest-analyzer.xhtml#logurl=https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/bsMZKyVeQDW4OXTWtgjfkA/runs/0/artifacts/public/logs/live_backing.log&only_show_unexpected=1

---

'''INFO'''

The test image is missing all the elements from the reference image.

'''INFO'''

Comment 1

[Intermittent Failures Robot]

2 failures in 3115 pushes (0.001 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 1
• mozilla-central: 1

Platform and build breakdown:

• windows10-aarch64-qr: 2
  • opt: 2

For more details, see:
https://treeherder.mozilla.org/intermittent-failures/bugdetails?bug=1718438&startday=2021-06-21&endday=2021-06-27&tree=all

Comment 2

[Intermittent Failures Robot]

7 failures in 723 pushes (0.01 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 7

Platform and build breakdown:

• windows10-aarch64-qr: 7
  • opt: 7

For more details, see:
https://treeherder.mozilla.org/intermittent-failures/bugdetails?bug=1718438&startday=2021-06-28&endday=2021-07-04&tree=all

Comment 3

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

Andrew, is there anything which can be done to reduce the failure rate of reftests on Windows AArch with WebRender enabled? The tasks ae failing permanently and contain several failures (not always the same).

Comment 4

[Intermittent Failures Robot]

12 failures in 2387 pushes (0.005 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 9
• mozilla-central: 2
• try: 1

Platform and build breakdown:

• windows10-aarch64-qr: 12
  • opt: 12

For more details, see:
https://treeherder.mozilla.org/intermittent-failures/bugdetails?bug=1718438&startday=2021-07-05&endday=2021-07-11&tree=all

Comment 5

[Intermittent Failures Robot]

22 failures in 2224 pushes (0.01 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• mozilla-beta: 5
• mozilla-central: 3
• autoland: 13
• try: 1

Platform and build breakdown:

• windows10-aarch64-qr: 22
  • opt: 22

For more details, see:
https://treeherder.mozilla.org/intermittent-failures/bugdetails?bug=1718438&startday=2021-07-12&endday=2021-07-18&tree=all

Comment 6

[Intermittent Failures Robot]

14 failures in 2940 pushes (0.005 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 5
• try: 1
• mozilla-beta: 5
• mozilla-central: 3

Platform and build breakdown:

• windows10-aarch64-qr: 14
  • opt: 14

For more details, see:
https://treeherder.mozilla.org/intermittent-failures/bugdetails?bug=1718438&startday=2021-07-19&endday=2021-07-25&tree=all

Comment 7

[Intermittent Failures Robot]

2 failures in 3131 pushes (0.001 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• mozilla-beta: 2

Platform and build breakdown:

• windows10-aarch64-qr: 2
  • opt: 2

For more details, see:
https://treeherder.mozilla.org/intermittent-failures/bugdetails?bug=1718438&startday=2021-07-26&endday=2021-08-01&tree=all

Comment 8

[Daniel Holbert [:dholbert]]

It looks like maybe this has gone away? No intermittents have been tagged with this bug for over a month. Most recent one is this one from July 28: https://treeherder.mozilla.org/logviewer?job_id=346634932&repo=mozilla-beta

Also: I think this is really a WebRender bug (or a failure being surfaced via WebRender).

In the failure log from comment 0, it looks like we're capturing & showing the completely wrong testcase's rendering for the "Test" screenshot; and this graphics error logging is in the log, just before the test failure:

INFO - [GFX1-]: EGL image is not valid.
INFO - [GFX1-]: Internal D3D11 error: HRESULT: 0x887A0005: Error allocating Texture2D
INFO - [GFX1-]: Context has been lost.
INFO - [GFX1]: Device reset due to WR device: 0x887a0006
INFO - [GFX1-]: GFX: RenderThread detected a device reset in PostUpdate
INFO - [GFX1]: Device reset due to WR device: 0x887a0006
INFO - [GFX1]: Device reset due to WR device: 0x887a0006
INFO - [GFX1]: Device reset due to WR device: 0x887a0006
INFO - REFTEST TEST-LOAD | file:///C:/tasks/task_1624770307/build/tests/reftest/tests/layout/reftests/bugs/76331-1-ref.html | 86 / 2048 (4%)
INFO - [GFX1-]: Failed to make render context current during destroying.
INFO - [GFX1-]: Failed to make render context current during destroying.
INFO - [GFX1-]: Failed to make render context current during destroying.

And in the most recent failure, there's a similar (but more concise) bit of graphics logging just before the test failure:

INFO - [GFX1-]: Readback took too long: 1273 ms
INFO - [GFX1]: Device reset due to WR device: 0x887a0006
INFO - [GFX1-]: GFX: RenderThread detected a device reset in PostUpdate
INFO - [GFX1]: Device reset due to WR device: 0x887a0006

The GFX: RenderThread detected a device reset in PostUpdate and Device reset due to WR device lines are present for both failures, so those seem like they're hinting at a root cause here.

--> Reclassifying this as WebRender, and adding those to the bug summary. (Though this may be WORKSFORME if we continue to not-see-new-reports.)