While we eventually plan to ship WebRender on NVIDIA devices with the proprietary drivers, for the moment, we need an initial group to help us validate Software WebRender.

I thought we wanted to do the experiment on all non accelerated Linux users with small screens?

(In reply to Jeff Muizelaar [:jrmuizel] from comment #1)

I thought we wanted to do the experiment on all non accelerated Linux users with small screens?

Yes you are right. I filed this while still waking up :).

This patch enables Software WebRender for all Linux users. If their
configuration is also allowlisted for (accelerated) WebRender, then they
will default to that over Software WebRender.

Backed out changeset 5d7d2da65d6f (bug 1677293) for gfx/layers/apz/test/mochitest failures.

Push with failures: https://treeherder.mozilla.org/jobs?repo=autoland&group_state=expanded&fromchange=a8d3345227775428e9ac826f09722851f6f4c3dc&searchStr=webrender%2Cmochitests&tochange=0cdbb584ff926da95d0fad97f53cc738301cc9df&selectedTaskRun=e4jbx3WMTdKq6wnv02FSOw.0

Backout link: https://hg.mozilla.org/integration/autoland/rev/0cdbb584ff926da95d0fad97f53cc738301cc9df

Failure log: https://treeherder.mozilla.org/logviewer?job_id=322093096&repo=autoland&lineNumber=3299

[task 2020-11-17T21:30:49.540Z] 21:30:49     INFO - TEST-START | gfx/layers/apz/test/mochitest/test_group_bug1464568.html
[task 2020-11-17T21:30:50.401Z] 21:30:50     INFO - GECKO(2559) | WaitUntilApzStable: flushed APZ repaints in parent proc, waiting for callback...
[task 2020-11-17T21:30:50.401Z] 21:30:50     INFO - GECKO(2559) | WaitUntilApzStable: APZ flush done in parent proc
[task 2020-11-17T21:30:50.421Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | WaitUntilApzStable: got apz-flush-done in child proc
[task 2020-11-17T21:30:50.436Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | WaitUntilApzStable: done promiseFocus
[task 2020-11-17T21:30:50.444Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | waiting for paint...
[task 2020-11-17T21:30:50.445Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | got MozAfterPaint: 0,0,0,0
[task 2020-11-17T21:30:50.446Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | Dispatching 1 onpaint listeners
[task 2020-11-17T21:30:50.453Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | got MozAfterPaint: 0,0,0,0
[task 2020-11-17T21:30:50.453Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | Dispatching 0 onpaint listeners
[task 2020-11-17T21:30:50.454Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | done...
[task 2020-11-17T21:30:50.455Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | WaitUntilApzStable: done promiseAllPaintsDone
[task 2020-11-17T21:30:50.459Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | PromiseApzRepaintsFlushed: Flushed APZ repaints, waiting for callback...
[task 2020-11-17T21:30:50.482Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | PromiseApzRepaintsFlushed: APZ flush done
[task 2020-11-17T21:30:50.489Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | WaitUntilApzStable: all done
[task 2020-11-17T21:30:50.496Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | waiting for paint...
[task 2020-11-17T21:30:50.537Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | got MozAfterPaint: 0,0,0,0
[task 2020-11-17T21:30:50.538Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | Dispatching 1 onpaint listeners
[task 2020-11-17T21:30:50.541Z] 21:30:50     INFO - GECKO(2559) | helper_bug1464568_transform.html | done...
[task 2020-11-17T21:30:50.550Z] 21:30:50     INFO - TEST-INFO | started process screentopng
[task 2020-11-17T21:30:50.980Z] 21:30:50     INFO - TEST-INFO | screentopng: exit 0
[task 2020-11-17T21:30:50.980Z] 21:30:50     INFO - Buffered messages logged at 21:30:49
[task 2020-11-17T21:30:50.980Z] 21:30:50     INFO - TEST-PASS | gfx/layers/apz/test/mochitest/test_group_bug1464568.html | Starting subtest helper_bug1464568_transform.html 
[task 2020-11-17T21:30:50.980Z] 21:30:50     INFO - Buffered messages finished
[task 2020-11-17T21:30:50.980Z] 21:30:50     INFO - TEST-UNEXPECTED-FAIL | gfx/layers/apz/test/mochitest/test_group_bug1464568.html | [Exception... "Component returned failure code: 0x80004001 (NS_ERROR_NOT_IMPLEMENTED) [nsIDOMWindowUtils.getOMTCTransform]"  nsresult: "0x80004001 (NS_ERROR_NOT_IMPLEMENTED)"  location: "JS frame :: resource://specialpowers/WrapPrivileged.jsm :: doApply :: line 118"  data: no] 
[task 2020-11-17T21:30:50.984Z] 21:30:50     INFO -     SimpleTest.ok@SimpleTest/SimpleTest.js:417:16
[task 2020-11-17T21:30:50.984Z] 21:30:50     INFO -     advanceSubtestExecutionWithFailure@gfx/layers/apz/test/mochitest/apz_test_utils.js:341:18
[task 2020-11-17T21:30:50.985Z] 21:30:50     INFO - TEST-PASS | gfx/layers/apz/test/mochitest/test_group_bug1464568.html | Starting subtest helper_bug1464568_opacity.html 
[task 2020-11-17T21:30:51.147Z] 21:30:51     INFO - GECKO(2559) | WaitUntilApzStable: flushed APZ repaints in parent proc, waiting for callback...
[task 2020-11-17T21:30:51.148Z] 21:30:51     INFO - GECKO(2559) | WaitUntilApzStable: APZ flush done in parent proc
[task 2020-11-17T21:30:51.164Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | WaitUntilApzStable: got apz-flush-done in child proc
[task 2020-11-17T21:30:51.165Z] 21:30:51     INFO - must wait for load
[task 2020-11-17T21:30:51.203Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | got MozAfterPaint: 0,0,0,0
[task 2020-11-17T21:30:51.204Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | Dispatching 0 onpaint listeners
[task 2020-11-17T21:30:51.205Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | WaitUntilApzStable: done promiseFocus
[task 2020-11-17T21:30:51.221Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | waiting for paint...
[task 2020-11-17T21:30:51.222Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | got MozAfterPaint: 0,0,0,0
[task 2020-11-17T21:30:51.223Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | Dispatching 1 onpaint listeners
[task 2020-11-17T21:30:51.226Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | done...
[task 2020-11-17T21:30:51.226Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | WaitUntilApzStable: done promiseAllPaintsDone
[task 2020-11-17T21:30:51.229Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | PromiseApzRepaintsFlushed: Flushed APZ repaints, waiting for callback...
[task 2020-11-17T21:30:51.246Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | PromiseApzRepaintsFlushed: APZ flush done
[task 2020-11-17T21:30:51.247Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | WaitUntilApzStable: all done
[task 2020-11-17T21:30:51.248Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | waiting for paint...
[task 2020-11-17T21:30:51.278Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | got MozAfterPaint: 0,0,0,0
[task 2020-11-17T21:30:51.278Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | Dispatching 1 onpaint listeners
[task 2020-11-17T21:30:51.279Z] 21:30:51     INFO - GECKO(2559) | helper_bug1464568_opacity.html | done...
[task 2020-11-17T21:30:51.280Z] 21:30:51     INFO - Not taking screenshot here: see the one that was previously logged
[task 2020-11-17T21:30:51.280Z] 21:30:51     INFO - TEST-UNEXPECTED-FAIL | gfx/layers/apz/test/mochitest/test_group_bug1464568.html | [Exception... "Component returned failure code: 0x80004001 (NS_ERROR_NOT_IMPLEMENTED) [nsIDOMWindowUtils.getOMTCTransform]"  nsresult: "0x80004001 (NS_ERROR_NOT_IMPLEMENTED)"  location: "JS frame :: resource://specialpowers/WrapPrivileged.jsm :: doApply :: line 118"  data: no] 
[task 2020-11-17T21:30:51.280Z] 21:30:51     INFO -     SimpleTest.ok@SimpleTest/SimpleTest.js:417:16
[task 2020-11-17T21:30:51.281Z] 21:30:51     INFO -     advanceSubtestExecutionWithFailure@gfx/layers/apz/test/mochitest/apz_test_utils.js:341:18
[task 2020-11-17T21:30:51.317Z] 21:30:51     INFO - GECKO(2559) | MEMORY STAT | vsize 2559MB | residentFast 165MB | heapAllocated 15MB
[task 2020-11-17T21:30:51.438Z] 21:30:51     INFO - TEST-OK | gfx/layers/apz/test/mochitest/test_group_bug1464568.html | took 1896ms
...
...
...

MOZ_WEBRENDER=1 => always hardware WebRender
MOZ_WEBRENDER=0 => always no WebRender
MOZ_WEBRENDER=1 + gfx.webrender.software = true => always software WebRender

I updated the patch, added more gtests for this, pending clean try:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=77896ff9454f06622bada0566a117223948ad698

Prior to landing I need to manually verify that each of these jobs runs with the right configurations (since the test might pass anyways).

The tests keep timing out, but an additional try with prints:

https://treeherder.mozilla.org/jobs?repo=try&revision=ba6df11f559b4301e0c6def398d53e4c4174dd0b

shows that we get no WR / basic, WR and SW WR when we expect. The change should now have no functional impact on CI so I don't think the time outs are related.

Backed out changeset 2991f291f95f (bug 1677293) for Wd failures at mozilla/webdriver/take_full_screenshot/iframe.py

Backout: https://hg.mozilla.org/integration/autoland/rev/ecd9813648e5697b0815e43cde11d277c20f5bb6

Failure push: https://treeherder.mozilla.org/jobs?repo=autoland&revision=2991f291f95fce56c0ada33d96091d5553c9bca5

Failure log: https://treeherder.mozilla.org/logviewer?job_id=322157663&repo=autoland&lineNumber=1517

I can confirm the failed job is running with Software WebRender. Investigation continues.

try with bug 1678044: https://treeherder.mozilla.org/jobs?repo=try&revision=5fcbc4c7121894133f09f9945e5af8c6745046e1

Backed out for GTest failure on GfxConfigManager

Backout link: https://hg.mozilla.org/integration/autoland/rev/feb67669cf8e2b52974cd987cc32c2d477d21155
Log link: https://treeherder.mozilla.org/logviewer?job_id=322390529&repo=autoland&lineNumber=28007

Bitten by own test affected by headless. It has gone full circle. This is a bad test, the patch is otherwise fine. Will update.

green try: https://treeherder.mozilla.org/jobs?repo=try&revision=532d28899f7da5ca2f98c82985f15eb9bcf82d52

Backed out changeset f6a72042dd56 (bug 1677293) for failures on fullscreen.py.

Push with failures: https://treeherder.mozilla.org/jobs?repo=autoland&group_state=expanded&selectedTaskRun=ZO8s90FwS7W3mZdVvBfQrw.0&fromchange=19bed19b33b1ee8fb274863d8b7815202c25f655&searchStr=wd&tochange=fd1683e51ec5eae6a5c5b516492d6a81eb06e7ea

Backout link: https://hg.mozilla.org/integration/autoland/rev/fd1683e51ec5eae6a5c5b516492d6a81eb06e7ea

Failures logs:
https://treeherder.mozilla.org/logviewer?job_id=322430038&repo=autoland&lineNumber=27855
https://treeherder.mozilla.org/logviewer?job_id=322433172&repo=autoland&lineNumber=19493
https://treeherder.mozilla.org/logviewer?job_id=322433171&repo=autoland&lineNumber=52984

And other /webdriver/ test failures.

Apparently the opt tests pass, but the debug ones don't. Last time I try to save CI resources for this bug ;).

The web driver tests seem to always pass for me with a local build. I'm trying an alternative now, roll out Software WebRender to a more limited set which hopefully excludes CI. I need to check both regular tests (running all linux64 debug tests) and some perf tests (because they use real hardware). Pending try to confirm this configuration doesn't change anything for CI.

https://hg.mozilla.org/mozilla-central/rev/f6a72042dd56

Backout merged to central: https://hg.mozilla.org/mozilla-central/rev/fd1683e51ec5eae6a5c5b516492d6a81eb06e7ea

https://hg.mozilla.org/mozilla-central/rev/fded65adecd4