Closed Bug 1440664 Opened 2 years ago Closed 2 years ago

Update webrender to 22b831c02479eea31821f49a0fac7dd699083557

Categories

(Core :: Graphics: WebRender, enhancement, P3)

60 Branch
enhancement

Tracking

()

RESOLVED FIXED
mozilla60
Tracking Status
firefox60 --- fixed

People

(Reporter: kats, Assigned: kats)

References

(Blocks 3 open bugs, )

Details

(Whiteboard: [gfx-noted])

Attachments

(6 files)

+++ This bug was initially created as a clone of Bug #1439565 +++

I'm filing this as a placeholder bug for the next webrender update. I may be running a cron script [1] that does try pushes with webrender update attempts, so that we can track build/test breakages introduced by webrender on a rolling basis. This bug will hold the try push links as well as dependencies filed for those breakages, so that we have a better idea going into the update of what needs fixing. I might abort the cron job because once things get too far out of sync it's hard to fully automate fixing all the breakages.

When we are ready to actually land the update, we can rename this bug and use it for the update, and then file a new bug for the next "future update".

[1] https://github.com/staktrace/moz-scripts/blob/master/try-latest-webrender.sh
Oh there's some new reftest failures. R1 failure is definitely from servo/webrender#2443. There's an R8 failure that shows up on debug only so far [1][2]. And there's an R5 debug intermittent [3] that looks like font/glyph raciness similar to what has been plaguing me on windows.

[1] https://treeherder.mozilla.org/logviewer.html#?job_id=163871539&repo=try&lineNumber=12925
[2] https://treeherder.mozilla.org/logviewer.html#?job_id=163908514&repo=try&lineNumber=12930
[3] https://treeherder.mozilla.org/logviewer.html#?job_id=163908500&repo=try&lineNumber=12898
The R5 debug thing seems to be bug 1238222 which is pre-existing, so I'm going to ignore it.
And the R8 failure seems to be bug 1439035 which might be qr-only but is also not new.
I did some bisection try pushes. This is the set of patches and try pushes I did at certain revisions:

| * | 69939456 Split DocumentMsg into SceneMsg and FrameMsg.
| * | c3fb6561 Address review comments.
| * | 272d8dbd Apply frame messages after building the scene in the sync path.
| * | be67702b Move the scene builder to its own file.
| * | 392c62b2 Rename prefix/postfix messages into scene/frame messages.
| * | b09c8e45 Keep track of the pending and built scenes separately on the render backend.
    -> https://treeherder.mozilla.org/#/jobs?repo=try&revision=d2c71afa2fd4046d5b5f32bc28a6ed8d8d13f2b7
       pretty broken
| * | edf064a4 Move dynamic properties into Document.
| * | 61bb29a4 Address some review comments.
| * | e39e0e7c Don't use the scene builder thread by default (yet).
| * | 5df631e1 Refactor transactions and forward them to the scene builder thread.
    -> https://treeherder.mozilla.org/#/jobs?repo=try&revision=1483f9ecc41a7da268534e5404b23eea25d5e3e2
       pretty broken, with panics
| * | c8ee8586 Handle removed pipelines.
| * | 4bd50525 Handle the pipeline epoch map properly.
| * | a481a202 Wake the render backend up after building the scene.
| * | a89270ac Move window_size from frame context to frame builder and update the clip scroll tree.
    -> https://treeherder.mozilla.org/#/jobs?repo=try&revision=926d271d51bc03defe5b5f65b835062a351252d2
       pretty broken, with panics
| * | 111cc480 Use Arc instead of Rc for clip chain nodes to make the clip scroll tree sendable.
| * | b4e030dc Begin patching the necessary information to the scene builder thread.
    -> https://treeherder.mozilla.org/#/jobs?repo=try&revision=eb964e91889b364dbdfa4f9e0a89f98a46da61d5
       build broken
| * | 5423ee16 Arc scene pipelines to get read access on both the render backend and scene builder threads.
    -> https://treeherder.mozilla.org/#/jobs?repo=try&revision=b98e2677246ebe330f12e319424b5d626c660997
       looks green (the reftest failures can be ignored, it's because this patch was based on a previous WR revision)
| * | 42b9bc2d Add the plumbing for a scene building thread.

So the problems seem to have started in 5423ee16..a89270ac but it might be a different problem what the final problem is at the end of the patchset.
WR @ 6d8609b79048384465863c1cd13e294fb26cae0f

https://treeherder.mozilla.org/#/jobs?repo=try&revision=d43989a894394b208494e5c888a5efe4ffb3c450
https://treeherder.mozilla.org/#/jobs?repo=try&revision=5f22c69ea1933b6d3f9a484ba858f7dbc5d460c4

Same

I found the problem with the async scene building patches, and put in a fix (servo/webrender#2484). Try push with current master + that fix:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=5ec349cb57128e974f9ef56b0474a8731cfcda2f
https://treeherder.mozilla.org/#/jobs?repo=try&revision=e3f9fe06d1dbd62d1d8856ac1f0a0ec4e36523b6

Looking pretty green so far, except for one R2 unexpected-pass. Looking back at the try pushes it looks like this used to be in R1, and before that it was in R6, showed up in the range 4e374f0c5139ca81ee8de99afe2fc3bc53e3632e..b53cf7799a0f5861dfa928ee0ccc66e8d0b78758 so probably from servo/webrender#2471.
Alias: wr-future-update
Assignee: nobody → bugmail
Summary: Future webrender update bug → Update webrender to 22b831c02479eea31821f49a0fac7dd699083557
Version: unspecified → 60 Branch
WR @ 22b831c02479eea31821f49a0fac7dd699083557

https://treeherder.mozilla.org/#/jobs?repo=try&revision=2e2f12e263c3482e5e0496e7a55ddd8be7d54550
https://treeherder.mozilla.org/#/jobs?repo=try&revision=2a04168f18cb26e477d46749a9bc68eba4ced5f3

Still in progress, but I'm expecting it to be green. I'll put the patches up but will wait for that try push before landing.
Comment on attachment 8955313 [details]
Bug 1440664 - Update gecko WR bindings for API changes in WR PR 2449.

https://reviewboard.mozilla.org/r/224464/#review230428
Attachment #8955313 - Flags: review?(bugmail) → review+
Comment on attachment 8955315 [details]
Bug 1440664 - Re-generate FFI header.

https://reviewboard.mozilla.org/r/224468/#review230442
Attachment #8955315 - Flags: review?(jmuizelaar) → review+
Comment on attachment 8955314 [details]
Bug 1440664 - Update Cargo lockfiles and re-vendor rust dependencies.

https://reviewboard.mozilla.org/r/224466/#review230440
Attachment #8955314 - Flags: review?(jmuizelaar) → review+
Comment on attachment 8955312 [details]
Bug 1440664 - Update webrender to commit 22b831c02479eea31821f49a0fac7dd699083557.

https://reviewboard.mozilla.org/r/224462/#review230438
Attachment #8955312 - Flags: review?(jmuizelaar) → review+
Comment on attachment 8955316 [details]
Bug 1440664 - Update reftest annotation for WR PR 2443.

https://reviewboard.mozilla.org/r/224470/#review230444
Attachment #8955316 - Flags: review?(jmuizelaar) → review+
Comment on attachment 8955317 [details]
Bug 1440664 - Update reftest annotation for changes in WR PR 2471.

https://reviewboard.mozilla.org/r/224472/#review230446
Attachment #8955317 - Flags: review?(jmuizelaar) → review+
Pushed by kgupta@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/6f54af02c2e6
Update webrender to commit 22b831c02479eea31821f49a0fac7dd699083557. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/c066c419e378
Update gecko WR bindings for API changes in WR PR 2449. r=kats
https://hg.mozilla.org/integration/autoland/rev/a5991299ebd0
Update Cargo lockfiles and re-vendor rust dependencies. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/6098d2c015d7
Re-generate FFI header. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/507a25c3afd0
Update reftest annotation for WR PR 2443. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/81f389c2524e
Update reftest annotation for changes in WR PR 2471. r=jrmuizel
Yeah I was noticing that too. I think I'll turn off the windows webgl QR tests (which I only enabled recently in bug 1426118) while we investigate, because I really don't want to back out this WR update. I'll just back out bug 1426118.
Flags: needinfo?(bugmail)
Depends on: 1442718
This boosted our performance:

== Change summary for alert #11897 (as of Thu, 01 Mar 2018 20:51:53 GMT) ==

Improvements:

  5%  Strings PerfHasRTLCharsRU osx-10-10 opt      982,541.38 -> 937,867.50
  3%  Strings PerfHasRTLCharsDE osx-10-10 opt      983,657.83 -> 952,768.42

For up to date results, see: https://treeherder.mozilla.org/perf.html#/alerts?id=11897
Doesn't make much sense, but that's OK.
Depends on: 1459760
If that's something that needs addressing, please file a new bug. Otherwise I'm just going to treat it as an interesting data point that I don't need to care about.
You need to log in before you can comment on or make changes to this bug.