All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=813996

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=b306904dcf81518604a6889ba0b455c121fdde61931719d5684948a3c49dd29a


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

android-nexus5
android-nexus7v2
android-one
android-webview-nexus6

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/11e5dd27840000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12943902440000

boliu: The bisect in #2 didn't complete, but it points to your CL, "aw: Make DrawFunctor event toplevel". Looks like most of the graphs recovered, is this already fixed?

Kicked off a second bisect just in case.

the DrawFunctor CL regressing webview is expected and reported many times already:  crbug.com/813972  essentially webview was missing a toplevel trace event which that CL fixed, so a bunch of cpu time was unaccounted for is now counted as webview cpu usage

that CL is in webview-only code though, so I don't see how it can possibly affect chrome though, so not sure about the non-webview regressions

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/12943902440000

Renderer observer of frame submission by jonross@chromium.org
https://chromium.googlesource.com/chromium/src/+/a2ff4f82109df55045dee9f54985a98054f86dc4

Omit the parent LayoutView's scroll offset from FrameRect by szager@chromium.org
https://chromium.googlesource.com/chromium/src/+/cf91f7965c29e2bc93e81f79cef5d77481b00e6e

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

For my change listed in #6 I have also since landed https://chromium-review.googlesource.com/c/chromium/src/+/934626 which should remove this impact in prod