All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=758219

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=a85897be348506c66944cbe26758d874ae2273995aa07936247d7a35949346fa


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

chromium-rel-mac11-pro

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8970458341272156208

=== Auto-CCing suspected CL author dmazzoni@chromium.org ===

Hi dmazzoni@chromium.org, 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 : Dominic Mazzoni
  Commit : 91ce4a5c047cf0a01f7d23c839bd4364428fa6ad
  Date   : Thu Aug 17 04:47:27 2017
  Subject: Add use counters for all ARIA attributes.

Bisect Details
  Configuration: mac_pro_perf_bisect
  Benchmark    : blink_perf.bindings
  Metric       : set-attribute-rare/set-attribute-rare
  Change       : 12.92% | 788.630668685 -> 686.721942751

Revision             Result                  N
chromium@495049      788.631 +- 7.39741      6      good
chromium@495073      787.827 +- 43.379       6      good
chromium@495075      797.511 +- 4.36357      6      good
chromium@495076      733.206 +- 6.64231      6      bad       <--
chromium@495079      731.951 +- 6.43327      6      bad
chromium@495085      728.725 +- 8.48136      6      bad
chromium@495096      686.722 +- 5.74705      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.bindings

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8970458341272156208


For feedback, file a bug with component Speed>Bisection

Code review for the relevant change:

https://chromium-review.googlesource.com/c/chromium/src/+/614642

 Issue 758226  has been merged into this issue.

 Issue 758227  has been merged into this issue.

 Issue 758216  has been merged into this issue.

 Issue 758220  has been merged into this issue.

I can reproduce locally. Just skipping the namespace and shadow
check doesn't help, but if I combine it with the event name
lookup above I can make the whole function faster.

Stay tuned, I'll try to upload a good fix.

Here's my proposed fix:

https://chromium-review.googlesource.com/c/chromium/src/+/633915

Feedback / drive-by reviews welcome.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/023904779abd234c473cc04f7d6e008715fc839f

commit 023904779abd234c473cc04f7d6e008715fc839f
Author: Dominic Mazzoni <dmazzoni@chromium.org>
Date: Fri Aug 25 05:20:35 2017

Optimize HTMLElement::ParseAttribute

Revision r495076 added some more use counters to HTMLElement::ParseAttribute,
but they regressed some performance benchmarks.

Optimize it by merging 6 conditionals and 2 hash lookups into a single hash
lookup that maps to a struct containing 3 possible triggers for any parsed
attribute:

* An event name
* A WebFeature for a use counter
* A HTMLElement member function to call

Local benchmarks show that with this change, the set-attribute-rare benchmark
is even faster than before r495076 landed.

Bug:  758219 
Change-Id: I7b4333be8360efaf7eda73a0be4fefc56a8a8a58
Reviewed-on: https://chromium-review.googlesource.com/633915
Reviewed-by: Kent Tamura <tkent@chromium.org>
Commit-Queue: Dominic Mazzoni <dmazzoni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#497327}
[modify] https://crrev.com/023904779abd234c473cc04f7d6e008715fc839f/third_party/WebKit/Source/core/html/HTMLElement.cpp
[modify] https://crrev.com/023904779abd234c473cc04f7d6e008715fc839f/third_party/WebKit/Source/core/html/HTMLElement.h

 Issue 759731  has been merged into this issue.