Bug 923881 - firefox, flash-plugin: Process /usr/lib/xulrunner/plugin-container was killed by signal 4
Summary: firefox, flash-plugin: Process /usr/lib/xulrunner/plugin-container was killed...
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Fedora
Classification: Fedora
Component: xulrunner
Version: 18
Hardware: i686
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
Assignee: Martin Stransky
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2013-03-20 16:36 UTC by rhbugz.domain.grmblfjx
Modified: 2013-05-06 17:43 UTC (History)
6 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2013-03-27 15:28:59 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)
crash log 1 (53.64 KB, text/plain)
2013-03-23 10:30 UTC, rhbugz.domain.grmblfjx
no flags Details
Okay (9.44 KB, image/jpeg)
2013-03-27 17:44 UTC, rhbugz.domain.grmblfjx
no flags Details

Description rhbugz.domain.grmblfjx 2013-03-20 16:36:44 UTC
Description of problem:
On this PC flash-plugin does not work. Every time Firefox encounters flash content the above error appears, no flash content works. I waited for new versions of firefox, xulrunner, flash-plugin, nothing changed.

Version-Release number of selected component (if applicable):
firefox.i686 19.0.2-1.fc18
xulrunner.i686 19.0.2-1.fc18
flash-plugin.i386 11.2.202.275-release

How reproducible:
every time

Steps to Reproduce:
1. start firefox
2. go to page with flash content
  
Actual results:
xulrunner crash, no flash content runs

Expected results:
flash content runs

Additional info:
Fedora 18 KDE Spin, 32bit
I submitted the crash report via abrt many times, but I don't know where this reports end up. If it helps, it says "uReport: BTHASH=a16f0a69955e0a1227d2dbc4d1fa4cb68d285be0" (I had to retype this by hand since there is no way to copy this from abrt m)

Comment 1 Martin Stransky 2013-03-22 10:21:14 UTC
Please attach a backtrace of the crash - http://fedoraproject.org/wiki/Debugging_guidelines_for_Mozilla_products#Application_crash

Comment 2 rhbugz.domain.grmblfjx 2013-03-23 10:30:30 UTC
Created attachment 715024 [details]
crash log 1

I followed the steps on the page. As firefox itself doesn't crash, I hit Ctrl-C to get to the (gdb) prompt after the xulrunner crash before entering the gdb commands. I don't know if this crash log is helpful, sorry.
Dekuju Martin.

Comment 3 Martin Stransky 2013-03-25 08:19:16 UTC
Hi,

yes, we need the firefox crash. We don't care about crashes in flash-plugin itself as fas as it's shielded by plugin-container. We only have to fix a case when whole browser crashes because of a plugin-container/flash crash.

Please try to carch the firefox crash if it's possible. Otherwise we close this report.

Comment 4 rhbugz.domain.grmblfjx 2013-03-25 12:05:15 UTC
Hi Martin
Firefox does not crash, only xulrunner. If I expressed myself unclear, please point out where.
ciao, Johannes

Comment 5 Martin Stransky 2013-03-27 15:00:23 UTC
xulrunner is just a part of firefox. So if firefox crashes we need the backtrace...otherwise we don't care.

Comment 6 rhbugz.domain.grmblfjx 2013-03-27 15:04:12 UTC
Please provide me with info on how to obtain the crashlog for xulrunner.

Comment 7 Martin Stransky 2013-03-27 15:28:59 UTC
The crash in xulrunner(plugin-container) happens in flash-plugin which is a closed source software. Even if you get the backtrace it does not help because we can't fix the flash plugin itself and there's any debuginfo for the flash-plugin.

Comment 8 rhbugz.domain.grmblfjx 2013-03-27 17:44:45 UTC
Created attachment 717186 [details]
Okay

Comment 9 Hironobu Ishii 2013-05-05 22:34:37 UTC
I experienced the same issue. My PC is Pentium 3 800MHz with 1GB mem.
"killed by signal 4" means plugin-container thought flash-plugin became no response.

I found following URLs.
1) http://www.tjdownes.com/post.cfm/fix-firefox-plugin-container-crash-when-debugging-flash
2) http://support.mozilla.org/pa-IN/questions/705281

Changing "dom.ipc.plugins.timeoutSecs" to 180(sec) works for me.
1st URL explains how to change this value.

My environment is:
$ LANG=C rpm -q firefox xulrunner flash-plugin
firefox-20.0-1.fc18.i686
xulrunner-20.0-1.fc18.i686
flash-plugin-11.2.202.280-release.i386

$ cat /proc/cpuinfo 
processor	: 0
vendor_id	: GenuineIntel
cpu family	: 6
model		: 8
model name	: Pentium III (Coppermine)
stepping	: 3
microcode	: 0x13
cpu MHz		: 801.803
cache size	: 256 KB
 ...
bogomips	: 1603.60

$ free
             total       used       free     shared    buffers     cached
Mem:       1027352     957036      70316          0      56148     485836
-/+ buffers/cache:     415052     612300
Swap:      2064380       4504    2059876

Comment 10 Hironobu Ishii 2013-05-05 23:55:48 UTC
Sorry, I made a mistake.
"signal 4" is SIGILL, not SIGKILL.
Extending timeout worked in some site, but it did not solve all cases.

Sorry for the noise.

Comment 11 Hironobu Ishii 2013-05-06 17:11:27 UTC
I found the cause.
Recent flash plugin uses SSE2 instructions.
And my PC does not support it.
Therefore, it is killed by SIGILL.

(gdb) info reg eip
eip            0xb553f380	0xb553f380
(gdb) disas /r 0xb553f380,+10
Dump of assembler code from 0xb553f380 to 0xb553f38a:
=> 0xb553f380:	0f 5b c1	cvtdq2ps %xmm1,%xmm0
   0xb553f383:	66 0f fe cb	paddd  %xmm3,%xmm1
   0xb553f387:	0f 29 45 d8	movaps %xmm0,-0x28(%ebp)
End of assembler dump.
(gdb) 

Thanks.

Comment 12 rhbugz.domain.grmblfjx 2013-05-06 17:43:26 UTC
Thanks for sharing your insights Ishii, it was very interesting.
However I don't have the same problem, as the machne in question is a Pentium 4 which I believe has SSE2.

btw, I also had a computer with the exact same processor back in the days

greetings


Note You need to log in before you can comment on or make changes to this bug.