tkent@ does it make sense to check the detail() == 1 (this field on MouseEvent  represents the number of clicks)

Specifically in this line of code:
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit/Source/core/html/HTMLAnchorElement.cpp&sq=package:chromium&type=cs&l=376&q=isLinkCLick

This would avoid the double invocation.

It sounds a good idea.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/e46f8fbfd5d2d3408d09d4262fb0f4f99c93a98f

commit e46f8fbfd5d2d3408d09d4262fb0f4f99c93a98f
Author: dtapuska <dtapuska@chromium.org>
Date: Thu May 26 04:10:53 2016

Only allow link clicks on single click.

When double clicking a link it would cause a double load. Ensure
link click only occurs on when the detail is 1.

BUG= 614467 

Review-Url: https://codereview.chromium.org/2011653003
Cr-Commit-Position: refs/heads/master@{#396116}

[add] https://crrev.com/e46f8fbfd5d2d3408d09d4262fb0f4f99c93a98f/third_party/WebKit/LayoutTests/fast/html/a-click.html
[modify] https://crrev.com/e46f8fbfd5d2d3408d09d4262fb0f4f99c93a98f/third_party/WebKit/LayoutTests/fast/loader/repeat-same-document-navigation.html
[modify] https://crrev.com/e46f8fbfd5d2d3408d09d4262fb0f4f99c93a98f/third_party/WebKit/Source/core/html/HTMLAnchorElement.cpp

Thanks for the prompt fix!!

If you can post what metrics you were querying specifically; like how did you determine 100ms timeout were caused by this.

The metrics were not gathered in UMA. ResourceDispatcher::Cancel was instrumented to dump without crashing whenever it asks the ResourceDispatcherHost to cancel a requet that had only lived < 100ms.

The logs were aggregated to find most used code paths in the wild. A lot came from click handlers on anchor elements, which I guessed to be either from these double clicks, or "enter clicks" where you hold enter and click a link. See  crbug.com/607361 .