Bug 83022 - memory leak
Summary: memory leak
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/Gallium/r600 (show other bugs)
Version: git
Hardware: Other All
: medium normal
Assignee: Default DRI bug account
QA Contact:
Depends on:
Reported: 2014-08-24 21:02 UTC by pavol
Modified: 2019-09-18 19:17 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:

Xorg.0.log (48.14 KB, text/plain)
2014-08-25 07:59 UTC, pavol
dmesg (78.17 KB, text/plain)
2014-08-25 07:59 UTC, pavol
glxinfo (59.26 KB, text/plain)
2014-08-25 08:00 UTC, pavol

Description pavol 2014-08-24 21:02:50 UTC
I am using VisIt to produce movies, however it seems that each state file leak memory and I cannot process all of them. I have tested it on nvidia workstation with proprietary drivers without problem.

Can you point me how to identify if it is a driver's problem or VisIt's?
Comment 1 Michel Dänzer 2014-08-25 01:45:18 UTC
First of all, please attach /var/log/Xorg.0.log and the output of dmesg and glxinfo.

The output from running the app with valgrind --leak-check=full (preferably with debugging symbols available for r600_dri.so) should be a good start to identify where the leaks are coming from.
Comment 2 pavol 2014-08-25 07:59:12 UTC
Created attachment 105219 [details]
Comment 3 pavol 2014-08-25 07:59:54 UTC
Created attachment 105220 [details]
Comment 4 pavol 2014-08-25 08:00:32 UTC
Created attachment 105221 [details]
Comment 5 pavol 2014-08-25 08:28:49 UTC
I tried to run it with valgrind and there are so many errors and it just crashed whole system (I mean disk IO was so high I had to press shutdown button, I do not get this, as I had no swap enabled). Second time I tried it and killed it just before system run out of resources. It looked like system works, however, Xorg process used all CPU (driver crashed and switched to SW rendering?) even that glxinfo reported it is running on R600.

On Intel driver it works well (I have dual system).
Comment 6 Michel Dänzer 2014-08-26 03:12:00 UTC
Can you try with fewer state files? The memory leaks should still be visible in the valgrind output.

Also, it would be better if you can exit VisIt cleanly instead of killing it, otherwise valgrind might incorrectly report memory as leaked which is normally freed correctly.
Comment 7 GitLab Migration User 2019-09-18 19:17:02 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/522.

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.