New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 767394 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Oct 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Windows
Pri: 1
Type: Bug-Regression



Sign in to add a comment

Regression:Black patch appear after clicking on reload button.

Reported by vku...@etouch.net, Sep 21 2017

Issue description

Chrome Version:63.0.3221.0 Revision e34bc3ba4a14b65c94f4b0ae103473934e32cfdb-refs/heads/master@{#503308}(32/64 Bit)  
OS:Windows (7,8,8.1,10),Linux (14.04 LTS)

What steps will reproduce the problem?
(1)Launch chrome and open NTP kill the page manually via chrome://kill
(2)Now reload the page and observe


Actual: Black patch appear after clicking on reload button.

Expected: Black patch should not be seen after clicking on reload button.

This is a regression  issue broken in 'M63' and below is the manual regression range
Good Build: 63.0.3216.0
Bad Build: 63.0.3217.0


 
Actual_Ntp.mp4
341 KB View Download

Comment 1 by vku...@etouch.net, Sep 21 2017

Labels: hasbisect-per-revision
Owner: clamy@chromium.org
Status: Assigned (was: Unconfirmed)
Good Build: 63.0.3216.0 (Revision : 502109)
Bad Build:  63.0.3217.0 (Revision : 502516)

You are probably looking for a change made after 502250 (known good), but no later than 502251 (first known bad).

CHANGELOG URL:
https://chromium.googlesource.com/chromium/src/+log/282ea833bef8dbf11acfa618d06568a910242c4d..0dbded05ad0ba1c3d8388fd649130509ab97f07a

Suspect: https://chromium.googlesource.com/chromium/src/+/0dbded05ad0ba1c3d8388fd649130509ab97f07a

Could you please look into the issue, pardon me if it has nothing to do with your changes and if possible please assign it to concern owner.

Note: Issue not seen on Mac OS.
Expected_Ntp.mp4
356 KB View Download

Comment 2 by treib@chromium.org, Sep 21 2017

Cc: jam@chromium.org nasko@chromium.org
Components: -UI>Browser>NewTabPage UI>Browser>Navigation Internals>Sandbox>SiteIsolation
Not related to the NTP, the same happens on google.com.
Labels: ReleaseBlock-Stable
Adding RB Label as this is recent Regression. Please remove if not required.
Thank You.
Labels: Proj-PlzNavigate
Owner: nasko@chromium.org
nasko@ - can you help with triage please? (reassigning since clamy@ is OOO)

This does indeed seem related to PlzNavigate - I can repro at ~ToT (r502380) (and so with PlzNavigate on by default / after r502251), but I cannot repro with --disable-browser-side-navigation switch.
Cc: kenrb@chromium.org lfg@chromium.org
<+kenrb@ and lfg@ who might have experience with similar issues that surfaced during the Site Isolation work>
Cc: clamy@chromium.org
I am not sure why this bug is marked as ReleaseBlock-Stable:

- IMO the fact that this is a recent regression (raised in #c3) doesn't seem (on its own) sufficient to declare this a blocker

- The only repro steps known today involve a renderer kill/crash.  I think this means the bug reproes rarely in practice.

- AFAICT the impact of the bug is purely aesthetic / non-functional (and it the impact lasts for ~1 second)


Please note that PlzNavigate is currently enabled at 100% for the stable channel.

Comment 7 by creis@chromium.org, Sep 22 2017

Cc: creis@chromium.org
Labels: -ReleaseBlock-Stable
I agree, this doesn't seem severe enough to be a release blocker, as noted in comment 6.

lukasza@: Does it repro on M61 or M62 with PlzNavigate enabled?  Or is it actually a regression in M63?  If it's the latter, maybe we can bisect to find what changed (i.e., run the bisect with PlzNavigate enabled).
This doesn't seem to be a recent regression for PlzNavigate - I can also repro the symptoms in 61.0.3163.100 (Linux, 64-bit) when launching it with --enable-browser-side-navigation flag.
Cc: ahemery@chromium.org

Comment 10 by nasko@chromium.org, Oct 16 2017

Owner: clamy@chromium.org
I'm reassigning this to clamy@, as I won't be able to work on this in the near future.
Kicked off investigations by introducing a WebContentsObserver and logging a number of potentially useful events (frame/navigation/some document stuff).

https://ahemery.users.x20web.corp.google.com/general_plznav_bugs/black_flashing_screen_after_crash_reload/web_content_observer_logs_and_timing
It looks like it might be a problem in the cc SetDeferCommits.
Without plz nav it is triggered from content::RenderFrameHostManager::Navigate.
With plz nav it is triggered (too late ?) from content::RenderFrameHostManager::DidNavigateFrame.
Project Member

Comment 13 by bugdroid1@chromium.org, Oct 20 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/b320a1bb07a05c28d56a0b80e307d3dfdc169ca9

commit b320a1bb07a05c28d56a0b80e307d3dfdc169ca9
Author: Arthur Hemery <ahemery@chromium.org>
Date: Fri Oct 20 13:53:26 2017

PlzNav: Moved visibility update in renderFrameManager.

The call that triggered the widget to be displayed in a reload was
placed in RenderFrameHostManager::CommitPendingIfNecessary which was
not reached in all cases (after a crash for example). Moved it to
RenderFrameHostManager::GetFrameHostForNavigation which is always
called.

Bug:  767394 
Cq-Include-Trybots: master.tryserver.chromium.linux:linux_site_isolation
Change-Id: Ifbbc54af972c6b7d36d28420b35f3743f1d6e568
Reviewed-on: https://chromium-review.googlesource.com/730009
Reviewed-by: Camille Lamy <clamy@chromium.org>
Commit-Queue: Arthur Hemery <ahemery@chromium.org>
Cr-Commit-Position: refs/heads/master@{#510423}
[modify] https://crrev.com/b320a1bb07a05c28d56a0b80e307d3dfdc169ca9/content/browser/frame_host/render_frame_host_manager.cc

Status: Fixed (was: Assigned)

Comment 15 by vku...@etouch.net, Oct 23 2017

Labels: TE-Verified-M64 TE-Verified-64.0.3247.0
Retested above issue on Windows(7,8,10) & Linux(14.04 LTS) OS using latest canary #64.0.3247.0 build and issue seems fixed hence adding TE-Verified labels.

Actual_Canary.mp4
384 KB View Download

Sign in to add a comment