Bug 567056 - [abrt] crash in ghostscript-8.71-1.fc12: Process /usr/bin/gs was killed by signal 11 (SIGSEGV)
Summary: [abrt] crash in ghostscript-8.71-1.fc12: Process /usr/bin/gs was killed by si...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Fedora
Classification: Fedora
Component: ghostscript
Version: 12
Hardware: i686
OS: Linux
low
medium
Target Milestone: ---
Assignee: Tim Waugh
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard: abrt_hash:f9ea2b85aa8d3b0d9caca0079ed...
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2010-02-21 10:20 UTC by Felix Möller
Modified: 2010-03-01 15:10 UTC (History)
5 users (show)

Fixed In Version: ghostscript-8.71-4.fc12
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2010-03-01 15:10:22 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
File: backtrace (364 bytes, text/plain)
2010-02-21 10:20 UTC, Felix Möller
no flags Details

Description Felix Möller 2010-02-21 10:20:48 UTC
abrt 1.0.7 detected a crash.

architecture: i686
cmdline: gs InfoInt_17_Containment_LaV.pdf
component: ghostscript
executable: /usr/bin/gs
kernel: 2.6.31.12-174.2.3.fc12.i686.PAE
package: ghostscript-8.71-1.fc12
rating: 4
reason: Process /usr/bin/gs was killed by signal 11 (SIGSEGV)
release: Fedora release 12 (Constantine)

backtrace
-----
Core was generated by `gs InfoInt_17_Containment_LaV.pdf'.
Program terminated with signal 11, Segmentation fault.
#0  0x00000002 in ?? ()

Thread 1 (Thread 6501):
#0  0x00000002 in ?? ()
No symbol table info available.
No shared libraries loaded at this time.
No symbol table is loaded.  Use the "file" command.
No symbol table is loaded.  Use the "file" command.

How to reproduce
-----
Just call gs from the commandline and it will crash imideatly.

dolphin seems to use gs to generate previews and therefore causes this crash a lot of times.

Sadly even running from gdb, the backtrace does not look any better ...

[fm@thinkpad ~]$ gdb gs
GNU gdb (GDB) Fedora (7.0.1-31.fc12)
Copyright (C) 2009 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "i686-redhat-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/bin/gs...Reading symbols from /usr/lib/debug/usr/bin/gs.debug...done.
done.
(gdb) r
Starting program: /usr/bin/gs 

Program received signal SIGSEGV, Segmentation fault.
0x00000001 in ?? ()

Comment 1 Felix Möller 2010-02-21 10:20:52 UTC
Created attachment 395334 [details]
File: backtrace

Comment 2 Jonathan Kamens 2010-02-21 15:24:39 UTC
Same problem here.  Even after running debuginfo-install ghostscript, the backtrace is still filled with garbage, so apparently whatever the bug is, it's trashing the stack.

Doing "yum downgrade ghostscript" followed by "ldconfig" seems to make the problem go away.

Comment 3 Jonathan Kamens 2010-02-21 15:43:16 UTC
I think ghostscript-8.71-4.fc12 fixes this.

Comment 4 Lester Caine 2010-02-22 09:07:28 UTC
Printing on my machine has been broken since the 19th ... and it would seem to be due to segfault in gs, but rolling back to 8.70 just gave black pages. How do I get 8.71.1.fc12 which the update service is providing replaced by 8.71.4.fc12 to see if this fixes the problem?

Comment 5 Tim Waugh 2010-02-22 09:26:09 UTC
It's here:
https://admin.fedoraproject.org/updates/ghostscript-8.71-4.fc12

Comment 6 Felix Möller 2010-02-22 09:43:40 UTC
This bug is in fact solved by ghostscript-8.71-4.fc12 for me.

Comment 7 Lester Caine 2010-02-22 10:18:39 UTC
What is more confusing is how the updates-testing repo became enabled on my machine in the first place. It's not something that *I* switched on, so why would these updates have suddenly started appearing on the machine?

Comment 8 Lester Caine 2010-02-22 11:33:18 UTC
OK
ghostscript-8.71-4.fc12 is at least printing pages, but there seems to be a problem with 'paging'. I've managed to get this mornings invoices off, which have footer elements so fill the page. When returning to do a set of labels which only half fill the page, then the printer stops - half page. I'm using the same printer configuration file on FC12 as an SUSE12 machine, and that gives the page feed. Also this was working fine up until Friday ... 
Is this another ghostscript regression, or should I be looking elsewhere to fix this problem?

Comment 9 Tim Waugh 2010-02-22 11:42:19 UTC
Hmm, not sure.  Please try ghostscript-8.70-2.fc12, which is the last known-good version-release:
  http://koji.fedoraproject.org/koji/buildinfo?buildID=148481

Does that fix it, and does updating to ghostscript-8.71-4.fc12 re-introduce the problem?

Comment 10 Lester Caine 2010-02-22 12:43:43 UTC
I found ghostscript-8.70-1.fc12 on the machine and that is doing the same thing ... so something else has changed. 
I have ghostscript-8.70-5.fc12 and ghostscript-8.71-1.fc12 appearing in the package manager. 70-5 prints black pages, as I had rolled back to that over the weekend. 
I've dropped back to Gutenprint 5.2.4-7 (from 5.2.5-2) but printing still stalls half page when the rest of the page is blank.
Where next?

Comment 11 Tim Waugh 2010-02-22 12:58:58 UTC
Lester: as this doesn't seem to due to the ghostscript update, please file a separate bug report against e.g. cups including the troubleshoot.txt output from the printing troubleshooter (System->Administration->Printing, then Help->Troubleshoot).  Thanks.

Comment 12 Tommi Kyntola 2010-02-23 11:20:49 UTC
I can verify that the 8.71.1.fc12 crashed on my x86_64 due to sigsegv. I noticed it using ps2pdf. Downgrade (to 70.5) worked and the 8.71-4 works now, too.


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