All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=830794

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=5d3c8da62254b81273650378ed28efbdf531828e497287da1e48e87f94f6a27c


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

linux-release

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/11ae98d4c40000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/11ae98d4c40000

Enable presentation feedback on Linux by penghuang@chromium.org
https://chromium.googlesource.com/chromium/src/+/400c1e296933da71f3b4e3a04b0843297dd4ffa4

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

I think enabling the presentation feedback shouldn't regress performance, but it changes the timing for firing the VSync callback. It could affect the some perf test results, if those results are based on the timestamp of VSync event.

Hi Sadrul & Ehan, Do you understand of those numbers? How those number are measured? 

I think we need to measure frame-times differently when presentation-feedback is used on Linux. We already do that on ChromeOS and Android, where we look at intervals between two consecutive 'DrmEventFlipComplete' and 'vsync_before' trace-events respectively, instead of 'BenchmarkInstrumentation::DisplayRenderingStats' events on other platforms [1]. I'll file a separate bug for that, and we should revisit this one once that is done to see if the numbers come back/stabilize or not, and whether there is additional work that needs to happen.

[1] https://cs.chromium.org/chromium/src/third_party/catapult/telemetry/telemetry/web_perf/metrics/rendering_stats.py?l=114

Reverting r548803 because it has been bisected as the root cause of several performance regressions.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/9c90ba20609cb897219c8b656b6255946c1b13d2

commit 9c90ba20609cb897219c8b656b6255946c1b13d2
Author: Annie Sullivan <sullivan@chromium.org>
Date: Wed Apr 11 00:27:07 2018

Revert "Enable presentation feedback on Linux"

This reverts commit 400c1e296933da71f3b4e3a04b0843297dd4ffa4.

Reason for revert: This CL caused several performance regressions.

Bug:  830776 , 830790 , 830792 , 830794 , 830802 , 830806 

Original change's description:
> Enable presentation feedback on Linux
> 
> Bug:  776877 
> Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
> Change-Id: I6faa89c42bb8a6bf43c4a9c6fe86826dca316e4f
> Reviewed-on: https://chromium-review.googlesource.com/998317
> Reviewed-by: Antoine Labour <piman@chromium.org>
> Commit-Queue: Peng Huang <penghuang@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#548803}

TBR=penghuang@chromium.org,piman@chromium.org

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

Bug:  776877 
Change-Id: I2f4d442305eb6505401e058773338b782c94c3ef
Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
Reviewed-on: https://chromium-review.googlesource.com/1005157
Reviewed-by: Annie Sullivan <sullivan@chromium.org>
Commit-Queue: Annie Sullivan <sullivan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#549693}
[modify] https://crrev.com/9c90ba20609cb897219c8b656b6255946c1b13d2/ui/gl/gl_surface_osmesa_x11.cc
[modify] https://crrev.com/9c90ba20609cb897219c8b656b6255946c1b13d2/ui/gl/gl_surface_osmesa_x11.h
[modify] https://crrev.com/9c90ba20609cb897219c8b656b6255946c1b13d2/ui/gl/gl_switches_util.cc

Marking Fixed as the CL was reverted. Please re-run the benchmark on perf trybots for linux before re-landing.