FS#47299 - [firefox] 42.0-2 crashes after high IO usage

Attached to Project: Arch Linux
Opened by Axilleas Pipinellis (markelos) - Tuesday, 08 December 2015, 18:03 GMT
Last edited by Jan de Groot (JGC) - Friday, 20 May 2016, 11:43 GMT
Task Type Bug Report
Category Packages: Extra
Status Closed
Assigned To Jan de Groot (JGC)
Evangelos Foutras (foutrelis)
Architecture x86_64
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

Description:

For a long time now, firefox (and thunderbird if that helps) reaches to a point where high IO is used rendering my system unresponsive and then it crashes.

Additional info:

* firefox 42.0-2
* linux 4.2.5-1

This on a Thinkpad x220 with 8GB of RAM, i7 CPU and a Samsung 840 PRO 256 SSD

Attached is the stack trace log from journald. I am running gdb now, but I see a lot of (no debugging symbols found) messages so I don't know if that will help.

Steps to reproduce:

I don't know whether this is reproducible, I just open firefox, work for some hours and then high IO until firefox crashes.
   firefox_stack_trace_journald.... (64.8 KiB)
This task depends upon

Closed by  Jan de Groot (JGC)
Friday, 20 May 2016, 11:43 GMT
Reason for closing:  Fixed
Comment by Axilleas Pipinellis (markelos) - Thursday, 10 December 2015, 17:29 GMT
The high IO seems to be caused by systemd-coredump, I noticed in iotop that when firefox is crashing the coredump reaches about ~100MB/s write IO.

* systemd 228-3
Comment by houmles (houmles) - Saturday, 19 December 2015, 18:27 GMT
This is very similar problem I am facing after last firefox major update (42 > 43) and I have even "better" HW specs so it's definitely not HW related.
You can easily reproduce it on website https://cryptowat.ch/okcoin/btccny ... switching options on this site or going fullscreen (F11) always crash whole firefox.
EDIT: forgot to add I was testing it on clear profile without any extensions and custom settings
This is what I got from gdb core dump:

Core was generated by `firefox'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x00007fefd4d3bc39 in raise () from /usr/lib/libpthread.so.0
[Current thread is 1 (Thread 0x7fefd512e740 (LWP 24605))]
(gdb) bt
#0 0x00007fefd4d3bc39 in raise () from /usr/lib/libpthread.so.0
#1 0x00007fefc6b5d260 in ?? () from /usr/lib/firefox/libxul.so
#2 0x00007fefc6dda6c5 in ?? () from /usr/lib/firefox/libxul.so
#3 <signal handler called>
#4 0x00007fefc6eba117 in ?? () from /usr/lib/firefox/libxul.so
#5 0x00007fefc761bda5 in JS::CompileOptions::CompileOptions(JSContext*, JSVersion) () from /usr/lib/firefox/libxul.so
#6 0x00007fefc76673ed in ?? () from /usr/lib/firefox/libxul.so
#7 0x00007fefc7666faf in ?? () from /usr/lib/firefox/libxul.so
#8 0x00007fefc74de360 in ?? () from /usr/lib/firefox/libxul.so
#9 0x00007fefc762d9fe in ?? () from /usr/lib/firefox/libxul.so
#10 0x00007fefc7419f1d in ?? () from /usr/lib/firefox/libxul.so
#11 0x00007fefc6e39ad7 in ?? () from /usr/lib/firefox/libxul.so
#12 0x00007fefc74ed4a9 in ?? () from /usr/lib/firefox/libxul.so
#13 0x00007fefc74ed162 in ?? () from /usr/lib/firefox/libxul.so
#14 0x00007fefc74ec928 in ?? () from /usr/lib/firefox/libxul.so
#15 0x00007fefc75a7527 in ?? () from /usr/lib/firefox/libxul.so
#16 0x00007fefc75a737e in ?? () from /usr/lib/firefox/libxul.so
#17 0x00007fefc75a7091 in ?? () from /usr/lib/firefox/libxul.so
#18 0x00007fefc75a6d78 in ?? () from /usr/lib/firefox/libxul.so
#19 0x00007fefd5076c81 in ?? ()
#20 0x00007fefb860b4e0 in ?? ()
#21 0x00007ffc42092f78 in ?? ()
#22 0x00007fefb860b420 in ?? ()
#23 0x00007fefb7b54910 in ?? ()
#24 0x0000000000000000 in ?? ()
Comment by Jan de Groot (JGC) - Wednesday, 13 April 2016, 12:57 GMT
Is this crash still valid with current versions of firefox?
Comment by houmles (houmles) - Wednesday, 13 April 2016, 20:38 GMT
Looks like it's fixed for me.
Just tested couple of webpages which always crashed the browser and it's working now.
Comment by Axilleas Pipinellis (markelos) - Saturday, 16 April 2016, 09:28 GMT
Seems fixed, thanks. I believe we can close this for now.

Loading...