Bug 29394 - Random segfaults from using the gdbus polkit branch
Summary: Random segfaults from using the gdbus polkit branch
Status: RESOLVED FIXED
Alias: None
Product: PolicyKit
Classification: Unclassified
Component: libpolkit (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: David Zeuthen (not reading bugmail)
QA Contact: David Zeuthen (not reading bugmail)
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-08-03 19:34 UTC by Petr Mrázek
Modified: 2011-02-23 05:56 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
Segfault backtrace (4.61 KB, application/octet-stream)
2010-08-03 19:35 UTC, Petr Mrázek
Details
Archive with many runs of the integtest tool (83.77 KB, application/x-gzip)
2010-08-04 13:48 UTC, Petr Mrázek
Details

Description Petr Mrázek 2010-08-03 19:34:46 UTC
I've rebuilt the whole stack of libraries (polkit-1, libpolkit-qt-1 and kdelibs) and now I'm getting some other error. The attached backtraces are from an application that just tests authorizations of many different actions, where some of them aren't known by the local authority. Polkit is used through two layers of extra libraries, so if it's not so useful, I'll cut it down to some smaller test case again.

The git revisions used:
polkit commit 90c0429f668be1cb675da26eba0ef505e6e3f123
glib commit 4bc9654c63feee596de5aa214d7cbd8fc293cd9d
Comment 1 Petr Mrázek 2010-08-03 19:35:38 UTC
Created attachment 37561 [details]
Segfault backtrace
Comment 2 David Zeuthen (not reading bugmail) 2010-08-04 09:42:02 UTC
Any chance you can run this with G_DBUS_DEBUG=all? Thanks
Comment 3 Petr Mrázek 2010-08-04 13:48:24 UTC
Created attachment 37577 [details]
Archive with many runs of the integtest tool

Ok, here it is. I did get two different segfaults (less common) and many assertion hits. Many runs also completed without any crashes.
Comment 4 David Zeuthen (not reading bugmail) 2010-08-05 06:30:13 UTC
(In reply to comment #3)
> Created an attachment (id=37577) [details]
> Archive with many runs of the integtest tool
> 
> Ok, here it is. I did get two different segfaults (less common) and many
> assertion hits. Many runs also completed without any crashes.

Ugh, never upload .tar.gz as attachments please - use plaintext attachments instead. That way it's much easier for people to read in a browser ... case in point: I wanted to read the attachments from my iPad last night but couldn't.
Comment 5 David Zeuthen (not reading bugmail) 2010-08-05 06:33:52 UTC
(In reply to comment #3)
> Created an attachment (id=37577) [details]
> Archive with many runs of the integtest tool
> 
> Ok, here it is. I did get two different segfaults (less common) and many
> assertion hits. Many runs also completed without any crashes.

Does valgrind say anything interesting about this?
Comment 6 Petr Mrázek 2010-08-13 04:37:44 UTC
(In reply to comment #5)
> Does valgrind say anything interesting about this?
Not really, just an invalid memory access to an unallocated address.
I updated to PolicyKit 0.97 along with rebuilding everything again and this seems to have vanished. The other bug with polkitd leaking and eventually stalling is back though.
Comment 7 David Zeuthen (not reading bugmail) 2010-08-13 07:52:33 UTC
(In reply to comment #6)
> (In reply to comment #5)
> > Does valgrind say anything interesting about this?
> Not really, just an invalid memory access to an unallocated address.
> I updated to PolicyKit 0.97 along with rebuilding everything again and this
> seems to have vanished. The other bug with polkitd leaking and eventually
> stalling is back though.

OK, I'll try to leak check + fix before releasing 0.98.
Comment 8 David Zeuthen (not reading bugmail) 2011-02-23 05:56:43 UTC
Closing as per comment 6.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.