Bug 1325454 - Latest firefox package, firefox-45.0.1-6.fc25, in rawhide gets segmentation faults all the time.
Summary: Latest firefox package, firefox-45.0.1-6.fc25, in rawhide gets segmentation ...
Keywords:
Status: CLOSED WORKSFORME
Alias: None
Product: Fedora
Classification: Fedora
Component: firefox
Version: rawhide
Hardware: Unspecified
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: 2016-04-08 21:00 UTC by stan
Modified: 2016-04-11 16:51 UTC (History)
5 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2016-04-11 16:51:13 UTC
Type: Bug


Attachments (Terms of Use)

Description stan 2016-04-08 21:00:17 UTC
Description of problem:
When running the latest firefox package in rawhide (fc25), the browser gets many segmentation faults.  One of them occurred while I was not doing anything.

Version-Release number of selected component (if applicable):
 firefox-45.0.1-6.fc25

How reproducible:
Seems to happen intermittently at anytime.  I didn't notice any activity that caused it consistently.

Steps to Reproduce:
1.  Install firefox-45.0.1-6.fc25 in rawhide
2.  Run as browser
3.

Actual results:
Pretty regular segmentation faults. (Five or six in half an hour?)

Expected results:
Runs as normal.

Additional info:
I think this might have something to do with gcc6.  I have been having problems compiling and running nightly with any optimization in rawhide since gcc6 became the default compiler version.  If I compile without optimization, -O0, nightly runs fine and is stable, but is slow, and on some things, like dns resolution, or communication with websites, very slow.  I have / had a bug open in the mozilla bugzilla, https://bugzilla.mozilla.org/show_bug.cgi?id=1256687. It's been duplicated to https://bugzilla.mozilla.org/show_bug.cgi?id=1245783. It seems there is an incompatibility between gcc6 optimization and mozilla firefox code.  It will get resolved, but it is painful in the meantime.

I've reverted to firefox-45.0.1-4.fc25, which has been stable for me since it was installed.

There is no information with the segfaults.  I run from a console so I can see error messages, and this is what prints.

$ /usr/bin/firefox
Segmentation fault (core dumped)

Comment 1 Martin Stransky 2016-04-11 11:25:07 UTC
Can you please attach a backtrace of the crash? You can diagnose the core dump or use ABRT to catch the crash.

Comment 2 stan 2016-04-11 16:51:13 UTC
This morning, I updated my system, and as part of that update, firefox updated to the -6 version again.  I have been running it since update, and there have been no seg faults.  In the time it has been running, there would have been at least three before.

So, I think something else in the update fixed whatever was causing the problem.  I noticed that gcc6 was updated.  Based on that, I am closing the bug.  If the problem recurs, I will re-open it, but I don't think that will happen.

And a good thing, because I didn't notice any abrt notices from the prior segfaults, and I'm not sure where to find core files.  I actually thought that they were being sent to /dev/null now.

Perhaps this implies that good things are going to happen in my nightly compiles in rawhide, also.  That would be great.


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