All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=809442

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=e5b495866dd1322d6fa31f7de159deadd63d27e6b8c4f7a31d7e776439b27164


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

android-webview-nexus6

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12ec1509840000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/12ec1509840000

Change webview SyncCompositor to use BeginFrameResponse messages. by dtapuska@chromium.org
chromium @ 9ef7db8d511a1ef590b0094954af949868989b30

Add unittests for MediaSesionTabHelper with fake LargeIconBridge by ssid@chromium.org
chromium @ a0c2a568819d862f5706a0bb07d96f9b153100f3

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

ssid's CL can't affect webview; it's all chrome-only code

dtapuska: can you check if this is real? I can't really imagine why your change would cause any regression either, but it's easy to miss things..

although looks like the graph is back down again when the memory leak was fixed. guess that must have been it