New issue
Advanced search Search tips

Issue 604106 link

Starred by 3 users

Issue metadata

Status: Fixed
Owner:
Closed: Apr 2016
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 1
Type: Bug

Blocking:
issue 602736



Sign in to add a comment

InitializeMojoIPCChannel failure in utility process

Project Member Reported by thestig@chromium.org, Apr 16 2016

Issue description

Version: r386529
OS: Windows for sure.

What steps will reproduce the problem?
(1) Run Chrome, or Chromium with --enable-cloud-print-proxy
(2) Go to chrome://devices
(3) Click Add printers
(4) Try to complete the cloud printer proxy set up.

What is the expected output?

Cloud print service process starts

What do you see instead?

Thread 0 CRASHED [EXCEPTION_BREAKPOINT @ 0x50596159 ] MAGIC SIGNATURE THREAD
	0x50596159	(chrome_child.dll -debugger_win.cc:21 )	base::debug::BreakDebugger()
	0x51bc3a63	(chrome_child.dll -child_thread_impl.cc:239 )	content::`anonymous namespace'::InitializeMojoIPCChannel
	0x51bc32be	(chrome_child.dll -child_thread_impl.cc:384 )	content::ChildThreadImpl::Init(content::ChildThreadImpl::Options const &)
	0x51bc293a	(chrome_child.dll -child_thread_impl.cc:328 )	content::ChildThreadImpl::ChildThreadImpl(content::ChildThreadImpl::Options const &)
	0x51f4dcb0	(chrome_child.dll -utility_thread_impl.cc:42 )	content::UtilityThreadImpl::UtilityThreadImpl()


Seen from users too:
Latest crash: f0863a2400000000 / 52.0.2709.1
Earliest crash: 2ed305a400000000 / 50.0.2638.0

Possible M-50 candidate too, but setting milestone to M-51 for now.
 
Note: For builds with DCHECK enabled, try to avoid  bug 429100  and bug 604077.
Though I have not been able to reproduce it with 49.x stable and 50.0.2661.75 stable on the same machine. Though I can with 52.0.2709.0 canary - b438112400000000.

As the cloud print proxy does not start, the printers I tried to add do not show up at https://www.google.com/cloudprint/#printers
s/does not start/encounters an error/ since it's an utility process that's crashing.
And in comment 1, I meant bug 429110.
Status: Started (was: Untriaged)
Hm. I suspect I know what's going on, and if it's true, it affects 51 as well (but only windows). Time to fire up my windows machine.
Oops. I meant to say, I think it started in 51, and 50 should be unaffected.
Project Member

Comment 7 by bugdroid1@chromium.org, Apr 18 2016

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d432f3e89f5d70c432810b9528d196629c4a50b8

commit d432f3e89f5d70c432810b9528d196629c4a50b8
Author: amistry <amistry@chromium.org>
Date: Mon Apr 18 21:15:45 2016

[mojo] Don't CHECK fail when a Mojo channel isn't available for a child process.

This is intended to be a mergable patch for M51.

BUG= 604106 

Review URL: https://codereview.chromium.org/1894933003

Cr-Commit-Position: refs/heads/master@{#388027}

[modify] https://crrev.com/d432f3e89f5d70c432810b9528d196629c4a50b8/content/child/child_thread_impl.cc

FWIW, with r388027, I no longer get the CHECK failure. However, after setting up the cloud print proxy, if I try to print to a proxied printer, I get a mysterious [Abort, Retry, Ignore] dialog. I'm not sure if that has anything to do with the Mojo work but just a heads up. There may be more P1 bugs coming your way.
When I built it locally, I was able to add the local printers and see them on the cloud print console. I didn't notice any other dialogs. Maybe it's a more recent change? I'll request a 51 merge and we'll see what happens in the beta branch.
Labels: Merge-Request-51
Oh, I re-read your comment. I didn't try printing. I'll take a quick look and see if I can repro myself.
Sure. For all I know, it couple be just me. I'm trying to track this down...
It's not just you. I'm seeing it as well.
That's good to know. BTW, I can't register cloud printers with Chromium snapshot builds. Probably due to the lack of some API keys. I guess I'll try some Chrome builds...
I filed  bug 604587  for the abort/retry/ignore error.
Blocking: 602736

Comment 17 by tin...@google.com, Apr 19 2016

Labels: -Merge-Request-51 Merge-Approved-51 Hotlist-Merge-Approved
Your change meets the bar and is auto-approved for M51 (branch: 2704)
We're cutting M51 beta candidate tomorrow, Wednesday @ 5:00 PM PST. Please merge your change to M51 branch 2704 before then. Thank you.
Project Member

Comment 19 by bugdroid1@chromium.org, Apr 20 2016

Labels: -merge-approved-51 merge-merged-2704
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/f40cefb712c4e8578f38edcf8f8795f86b525754

commit f40cefb712c4e8578f38edcf8f8795f86b525754
Author: Anand K. Mistry <amistry@chromium.org>
Date: Wed Apr 20 05:15:39 2016

[mojo] Don't CHECK fail when a Mojo channel isn't available for a child process.

This is intended to be a mergable patch for M51.

BUG= 604106 

Review URL: https://codereview.chromium.org/1894933003

Cr-Commit-Position: refs/heads/master@{#388027}
(cherry picked from commit d432f3e89f5d70c432810b9528d196629c4a50b8)

Review URL: https://codereview.chromium.org/1897823007 .

Cr-Commit-Position: refs/branch-heads/2704@{#141}
Cr-Branched-From: 6e53600def8f60d8c632fadc70d7c1939ccea347-refs/heads/master@{#386251}

[modify] https://crrev.com/f40cefb712c4e8578f38edcf8f8795f86b525754/content/child/child_thread_impl.cc

Status: Fixed (was: Started)
I verified the fix on 52.0.2712.0 canary before landing.
It looks like this might be the cause of P0 for M50 https://bugs.chromium.org/p/chromium/issues/detail?id=607677.
... but the problem code landed after the 50 branch point. crrev.com/eb199fbaecd8e5d294010dfbe8897b7d83aeab07 is the original landing, which is in  51.0.2664.0
Cc: thestig@chromium.org
 Issue 605949  has been merged into this issue.

Sign in to add a comment