Bug 633257 - Abrt generates wrong backtraces
Summary: Abrt generates wrong backtraces
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Fedora
Classification: Fedora
Component: abrt
Version: 14
Hardware: i686
OS: Linux
low
medium
Target Milestone: ---
Assignee: Karel Klíč
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard: abrt_hash:cbfef7a07a520455c6fd67acfaa...
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2010-09-13 12:37 UTC by Nils Philippsen
Modified: 2013-03-03 23:01 UTC (History)
9 users (show)

Fixed In Version:
Clone Of: 631674
Environment:
Last Closed: 2012-08-10 11:28:51 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)

Description Nils Philippsen 2010-09-13 12:37:15 UTC
I got this bug report with a dubius abrt-generated backtrace -- the flagged line (see below and in attachment #445888 [details]) shouldn't be able to cause a SIGSEGV.

Not sure what is wrong here, perhaps debuginfo not matching the executable/library?

+++ This bug was initially created as a clone of Bug #631674 +++

abrt version: 1.1.13
architecture: i686
Attached file: backtrace
cmdline: ufraw
comment: can supply an image if it would help (18mb in size)
component: ufraw
crash_function: _lf_detect_cpu_features
executable: /usr/bin/ufraw
kernel: 2.6.35.4-12.fc14.i686
package: ufraw-0.17-1.fc14
rating: 4
reason: Process /usr/bin/ufraw was killed by signal 11 (SIGSEGV)
release: Fedora release 14 (Laughlin)
time: 1283928439
uid: 500

How to reproduce
-----
1. cd to directory containing images (have not tested from different dir)
2. started ufraw as unpriveledged user from command line (gnome-terminal)
3. selected single RAW image (taken with canon 7d)
4. selected open-> crash

--- Additional comment from gareth.glaccum on 2010-09-08 02:51:18 EDT ---

Created an attachment (id=445888)
File: backtrace

--- Additional comment from nphilipp on 2010-09-13 08:33:50 EDT ---

Can you reproduce this error reliably? I'm asking because the the attached backtrace generated by abrt is surely broken, this code line ...:

	                    cpuflags |= LF_CPU_FLAG_MMX;

... shouldn't be able to cause a segmentation violation.

Comment 1 Jiri Moskovcak 2011-05-31 12:01:37 UTC
(In reply to comment #0)
> I got this bug report with a dubius abrt-generated backtrace -- the flagged
> line (see below and in attachment #445888 [details]) shouldn't be able to cause a
> SIGSEGV.
> 
> Not sure what is wrong here, perhaps debuginfo not matching the
> executable/library?
> 
- in that case I would expect a warning in GDB output and since it's not there I suspect abrt bt parser

Comment 2 Jiri Moskovcak 2012-08-10 11:28:51 UTC
There has been a lot of changes in the code since abrt-1.1.13 and the backtrace generation logic has been improved so I believe it's either fixed or doesn't happen so often...


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