The first test (DontReuse...) (which was added in the revision above) is also failing on the DrMemory full bot. But not the other test. Hmmmm.

I just saw https://github.com/DynamoRIO/drmemory/issues/1528.

Derek: is this something the feature owner should look into, or is it more likely something to do with DrMemory itself?

The cause of each test failure in https://github.com/DynamoRIO/drmemory/issues/1528 is unknown.  We expect most or all to be test problems: hardcoded timeouts or the like where the test just can't handle a heavyweight tool.  I.e., there are no known Dr. Memory bugs there.  But, the failures have not been analyzed.

I'm looking.  Alex and I were aware the DontReusePendingDeleteRenderViewHostForSubframe test might not exercise the desired path on very slow bots, but it seemed like it should still pass either way.  Instead, it appears to be failing:

c:\b\build\slave\drm-cr\build\src\content\browser\site_per_process_browsertest.cc(4199): error: Expected: (rvh_a_routing_id) != (rvh_a2->GetRoutingID()), actual: 1 vs 1

I don't know about the other two.  Both NestedPopupMenuTest (from Ken) and SequentialFocusNavigation (from Alex) time out frequently on other bots as well:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#testType=content_browsertests&tests=SitePerProcessBrowserTest.NestedPopupMenuTest
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#testType=content_browsertests&tests=SitePerProcessBrowserTest.SequentialFocusNavigation

I'm guessing we should have a separate bug for those, though I'm not sure why they started failing on DrMemory at the same time as my CL.

Fix for DontReusePendingDeleteRenderViewHostForSubframe is on the way in https://codereview.chromium.org/1851633002/.

I may have to ask Alex and Ken to look at the other two tests.  I'm not sure why they would start failing at the same time as mine, given that they've been flaky on other bots for a while (and don't seem particularly affected by my change at first glance).

For reference, they were added in https://codereview.chromium.org/1820043002 and https://codereview.chromium.org/1500873002.

OK ... the try bots stopped my test disabling patch https://codereview.chromium.org/1848743002, given a fix is coming I'll just close the patch.
 
Sounds like the other two tests should be disabled more generally. I'll do that today.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/05a1cb791558af2b5ef244ffbf2681958880f8ed

commit 05a1cb791558af2b5ef244ffbf2681958880f8ed
Author: creis <creis@chromium.org>
Date: Thu Mar 31 21:26:24 2016

Fix flaky DontReusePendingDeleteRenderViewHostForSubframe test.

BUG= 599344 
TEST=Test goes green on DrMemory bot.

Review URL: https://codereview.chromium.org/1851633002

Cr-Commit-Position: refs/heads/master@{#384396}

[modify] https://crrev.com/05a1cb791558af2b5ef244ffbf2681958880f8ed/content/browser/site_per_process_browsertest.cc

Looks like we can track the followup work for the other two tests in issue 599730.