346308 - CPU Spikes to 100%, Bon Echo Hangs [@ nsTextNode::Release] [@ nsRange::OwnerChildRemoved]

Status: VERIFIED DUPLICATE of bug 346233

(Core :: DOM: Core & HTML, defect)

Description

[John Liebson]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b1) Gecko/20060728 BonEcho/2.0b1
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b1) Gecko/20060728 BonEcho/2.0b1

Problem only arose with today's release. Constant hanging of BE, such as when I was trying to find any Bugzilla reports about the problem. I have to kill Bon Echo; when I restart BE and tell it to restore the session, it works properly.

This has happened perhaps ten or twelve times in the past hour to one and one-half hours.

Reproducible: Sometimes

Steps to Reproduce:
1. All you have to do is try to use today's release of Bon Echo. The failures will happen by themselves, whether you have one or more tabs open.
2.
3.

Actual Results:  
Bon Echo hangs; no crashes, thus no Talkback reports.

Expected Results:  
Bon Echo should work properly.

Please see messages in this thread:

http://forums.mozillazine.org/viewtopic.php?p=2396663#2396663

Comment 1

[Unknown W. Brackets]

I'm definitely seeing this too.  Seems random... but I'm seeing spikes to 50% constant usage, and Firefox not responding any further.

I can't seem to reproduce it.  Happened while using midas, happened when reloading a link a few times...

-[Unknown]

Comment 2

[Simon Bünzli]

Confirming. I've got the impression that it usually happens when either hovering over a link or just at the point of clicking one.

Comment 3

[Steve England [:stevee]]

Regression range:
2006-07-27-04 - works
2006-07-28-03 - hanging observed

http://tinderbox.mozilla.org/bonsai/cvsquery.cgi?treeid=default&module=PhoenixTinderbox&branch=MOZILLA_1_8_BRANCH&branchtype=match&dir=&file=&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=explicit&mindate=2006-07-27+09%3A57%3A26&maxdate=2006-07-28+02%3A57%3A21&cvsroot=%2Fcvsroot

Comment 4

[Hammer]

I confirm this also happening in Safe Mode.
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b1) Gecko/20060728 BonEcho/2.0b1 ID:2006072803

Comment 5

[Marcia Knous [:marcia]]

nominating.

Comment 6

[Andreas Morawietz]

I just used crash2 as firefox hung.

Please check this talkback id, I think there's a lot of information in it.

TB21538751Q

Comment 7

[Adam Guthrie]

See also bug 346233 which complains about a hang in nsRange.

Comment 8

[Peter Petrov]

I have the same problem with the 20060728 build. I used Sysinternals Process Explorer to check which thread actually consumes the 100% CPU time. It shows the thread as:

firefox.exe!jpeg_fdct_islow+0x27a62

Hope this helps.

Comment 9

[gene c]

Happens on linux too.
I believe this began on branch build sometime around 7/14/2006 or thereabouts.
Confirmed again on this build:

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1b1) Gecko/20060729 BonEcho/2.0b1 ID:2006072907

It always shows up for me when I leave the firefox running (today the _only_ tab open was to the mozillazine build forum) - I come back and it is in the 100% CPU state - and must be killed.

g/

Comment 10

[Nickolay_Ponomarev]

This is most likely a dupe of bug 346233, I posted some details there.

Comment 11

[gene c]

It seems to be worse if ff starts up in offering to restore a prev session. It happens either way, but I seem to be able to trigger it more easily in the former case.

Comment 12

[Bradley Chapman (not reading bugmail, still gone but not forever)]

I have four SIGSEGV-generated talkbacks for this hang; three of them have nsRange::PopRanges() in the fourth stack frame. When the hang happens again I'm going to try to get the program into a debugger and see what happens.

Comment 13

[Jonathan Stanley]

I noticed this as of a couple days ago thinking it was a compiler issue with my own spun builds, though having just had two hangs (in as many hours) with the following tinderbox build:

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b1) Gecko/20060730 BonEcho/2.0b1 ID:2006073012

... and seeing this bug which describes the same problem, this is another "me too".

Comment 14

[Dave Townsend [:mossop]]

*** Bug 346681 has been marked as a duplicate of this bug. ***

Comment 15

[Steve England [:stevee]]

*** Bug 346700 has been marked as a duplicate of this bug. ***

Comment 16

[gene c]

Attachment: gdb backtrace - debugger attached after FF hang with 100% CPU

When FF went to 100% CPU I attached gdb and did a backtrace.

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1b1) Gecko/20060729 BonEcho/2.0b1 ID:2006072907

Comment 17

[Robert O'Callahan (:roc) (email my personal email if necessary)]

That stack trace isn't useful, you need a debuginfo RPM.

Comment 18

[evilzenscientist]

Any update on this one - the current nightly is way too unstable for any more testing right now.

Any idea on a fix to test?

Comment 19

[David Miller]

Still seeing it here.

user-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b1) Gecko/20060803 BonEcho/2.0b1 ID:2006080305

Comment 20

[evilzenscientist]

(In reply to comment #19)
> Still seeing it here.
> 
> user-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b1)
> Gecko/20060803 BonEcho/2.0b1 ID:2006080305
> 

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b1) Gecko/20060803 BonEcho/2.0b1

Mine is a lot more stable today; I've not crashed in about 2.5 hours of use.

Anyone know if a fix was checked into the 0803 nightly?

Comment 21

[David Miller]

nevermind that.  it looks like it's a total different bug with similar symptoms.  sorry about that.

Comment 22

[mmortal03]

(In reply to comment #20)
> (In reply to comment #19)
> > Still seeing it here.
> > 
> > user-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b1)
> > Gecko/20060803 BonEcho/2.0b1 ID:2006080305
> > 
> 
> Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b1) Gecko/20060803
> BonEcho/2.0b1
> 
> Mine is a lot more stable today; I've not crashed in about 2.5 hours of use.
> 
> Anyone know if a fix was checked into the 0803 nightly?
> 

Yeah, the check-in for Bug 346233, I think.

Comment 23

[Steve England [:stevee]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8.1b1) Gecko/20060803 BonEcho/2.0b1 ID:2006080300
The random hanging has stopped for me too.

Comment 24

[Nickolay_Ponomarev]

*** This bug has been marked as a duplicate of 346233 ***

Comment 25

[Tracy Walker [:tracy]]

verified in fixed dupe