All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=722481

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgqtrAnQgM


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

android-nexus7v2

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8979504357144275344

=== Auto-CCing suspected CL author jl@opera.com ===

Hi jl@opera.com, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : jl
  Commit : 2a32710d5f0e31cf648120a9724d5c8b0d8e5a88
  Date   : Fri May 12 12:36:29 2017
  Subject: Inherit [Unforgeable] attributes between components

Bisect Details
  Configuration: android_nexus7_perf_bisect
  Benchmark    : media.android.tough_video_cases
  Metric       : vm_shared_dirty_delta/tulip2.m4a_gpu
  Change       : 50.00% | 16.0 -> 24.0

Revision             Result                  N
chromium@471271      16.0 +- 0.0             6      good
chromium@471278      16.6667 +- 3.65148      6      good
chromium@471282      16.0 +- 0.0             6      good
chromium@471283      16.0 +- 0.0             6      good
chromium@471284      24.0 +- 0.0             6      bad       <--
chromium@471285      24.0 +- 0.0             6      bad
chromium@471298      24.0 +- 0.0             6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests media.android.tough_video_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8979504357144275344

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5895538722996224


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

 Issue 722917  has been merged into this issue.

 Issue 722918  has been merged into this issue.

 Issue 722919  has been merged into this issue.

Adding reviewers. Does it make sense that this CL would eat some memory?

It doesn't make much sense to me that it would have a significant effect on memory usage. Of course, the regression seems to be a fairly small effect. Still, I don't really see how this commit could cause it.

These measurements seem to go up and down quite a lot (compared to the size of the regression.) How sure are we something out of the ordinary actually happened?

Seeing the graphs, it seems to me like we're very close to hit the threshold of memory page size, so the graphs are bumping up and down.  Except for crowd.ogg_gpu, the rest of graphs seem reasonably subtle.  I don't see a reason that we think this is a regression.

I close this issue as WontFix.  Please reopen this if anyone has any opinions.