Bug 91827 - gsraytrace fails with Input/Output error intel_do_flush_locked
Summary: gsraytrace fails with Input/Output error intel_do_flush_locked
Status: NEW
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: 11.0
Hardware: Other All
: medium normal
Assignee: Ian Romanick
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-08-31 19:36 UTC by Jonas Wielicki
Modified: 2018-09-04 15:57 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Jonas Wielicki 2015-08-31 19:36:04 UTC
The gsraytrace demo from mesademos 8.2 fails with the following output on mesa-11.0.0-rc2 on Intel Sandybridge:

    $ LD_LIBRARY_PATH=/usr/local/lib /usr/lib64/mesa/gsraytrace      
    GL_RENDERER = Mesa DRI Intel(R) Sandybridge Mobile 

    ESC                 = exit demo
    left mouse + drag   = rotate camera

    intel_do_flush_locked failed: Input/output error
    intel_do_flush_locked failed: Input/output error

No special input is required, the demo crashes right away. The same also happens with the system mesa which is at version 10.6.3.

intel GPU:
00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller (rev 09)
Comment 1 Albert Freeman 2015-09-01 13:09:08 UTC
On my system if the window size is lowered to 20 by 20 (or a similar or lower size (to use less GPU per "draw call"))). And Draw() is called directly a single time (instead of using glutMainLoop) (or more Draw()s for lower window sizes). Then the application stops causing a graphics reset and the "intel_do_flush_locked" error message disappears...
Comment 2 Albert Freeman 2015-09-01 13:11:13 UTC
Note: "draw call" is not the usual use of draw call in the above statement.
Comment 3 Albert Freeman 2015-09-01 13:22:46 UTC
Also, if I set the window size to something large like 750 by 750 and then perform two Draw()s then a graphics reset happens on the first Draw() but the intel error messages happen on the second Draw() (in other words: if the second Draw() is removed from the source code, the program terminates and no intel error messages are issued).
Comment 4 Albert Freeman 2015-09-01 13:38:12 UTC
In the second Draw() call, the function calls within which the error messages are issued are: glDrawArrays(GL_POINTS, 0, pg) and glutSwapBuffers().
Comment 5 asimiklit 2018-09-04 09:55:28 UTC
Could you please recheck this issue on latest mesa with commit?

1b0df8a46020cc88afeaa4decb42a782ab168afb

i965/gen6/xfb: handle case where transform feedback is not active
Comment 6 asimiklit 2018-09-04 10:23:10 UTC
Could you please recheck this issue on latest mesa with commit:

1b0df8a46020cc88afeaa4decb42a782ab168afb

i965/gen6/xfb: handle case where transform feedback is not active
Comment 7 Jonas Wielicki 2018-09-04 15:57:17 UTC
Sorry, I don't have the original hardware available any more to reproduce.

FWIW, the apitrace I took back then runs fine on my current hardware (Haswell-based though).


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.