New issue
Advanced search Search tips

Issue 880856 link

Starred by 1 user

Issue metadata

Status: Assigned
Owner:
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

19511.1% regression in browser_tests at 588691:588704

Project Member Reported by emir...@chromium.org, Sep 5

Issue description

Clear regression signal on all platforms.
 
All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=880856

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=adf521d67d314722394c2190f20b24a117137c2a981d3db9d81bc8ceca484160


Bot(s) for this bug's original alert(s):

chromium-webrtc-rel-linux
Owner: sunn...@chromium.org
Status: Assigned (was: Untriaged)
sunnyps@ can you PTAL as theis looks related to https://chromium.googlesource.com/chromium/src/+/c13dcf77ea89896271187aa929923ed84bc57dd0.
Project Member

Comment 3 by bugdroid1@chromium.org, Sep 14

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/2bbe289cc8e1ba48e6aac024521bafeb489f3ea0

commit 2bbe289cc8e1ba48e6aac024521bafeb489f3ea0
Author: Sunny Sachanandani <sunnyps@chromium.org>
Date: Fri Sep 14 00:25:21 2018

Revert "cc: Finish frame early if main thread is idle regardless of redraw state"

This reverts commit c13dcf77ea89896271187aa929923ed84bc57dd0.

Reason for revert: Cause of many regressions (see linked bugs)

Original change's description:
> cc: Finish frame early if main thread is idle regardless of redraw state
> 
> This condition was added originally in this patch:
> https://codereview.chromium.org/27200003
> 
> That patch was trying to trigger early deadlines when the compositor
> thread had work to do but the main thread had aborted.
> 
> In a world where we have begin frame acks even when frames aren't
> produced, it's worth triggering early deadlines even when the compositor
> thread doesn't have any work to do either so that the display compositor
> can go ahead and do work earlier.
> 
> This is intended as a short-term fix to solve smoothness issues with
> surface for video with 30fps videos.  Ideally the DisplayCompositor will
> have some followups as well so that one client that is slow to respond
> will not cause smoothness issues with another client.
> 
> Bug:  874676 
> Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;master.tryserver.blink:linux_trusty_blink_rel
> Change-Id: Ib442797dda715c7a34ef8c1cba740cb44c94eb60
> Reviewed-on: https://chromium-review.googlesource.com/1205101
> Reviewed-by: enne <enne@chromium.org>
> Commit-Queue: Sunny Sachanandani <sunnyps@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#588694}

TBR=enne@chromium.org,sunnyps@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug:  874676 , 880957, 880856
Change-Id: I7813aab2c71df33b0d8c184c174f2e73cbf64256
Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;master.tryserver.blink:linux_trusty_blink_rel
Reviewed-on: https://chromium-review.googlesource.com/1211664
Reviewed-by: Sunny Sachanandani <sunnyps@chromium.org>
Reviewed-by: enne <enne@chromium.org>
Commit-Queue: Sunny Sachanandani <sunnyps@chromium.org>
Cr-Commit-Position: refs/heads/master@{#591224}
[modify] https://crrev.com/2bbe289cc8e1ba48e6aac024521bafeb489f3ea0/cc/scheduler/scheduler_state_machine.cc
[modify] https://crrev.com/2bbe289cc8e1ba48e6aac024521bafeb489f3ea0/cc/scheduler/scheduler_state_machine_unittest.cc

Sign in to add a comment