Bug 96348 - [regression bisected] bloom demo rendering glitches
Summary: [regression bisected] bloom demo rendering glitches
Status: RESOLVED FIXED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/Vulkan/intel (show other bugs)
Version: git
Hardware: Other All
: medium normal
Assignee: Intel 3D Bugs Mailing List
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-06-02 23:19 UTC by Nanley Chery
Modified: 2016-09-23 22:26 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Expected result (484.91 KB, image/png)
2016-06-02 23:19 UTC, Nanley Chery
Details
Actual result (443.65 KB, image/png)
2016-06-02 23:19 UTC, Nanley Chery
Details
Actual result with synchronization fix (485.25 KB, image/png)
2016-06-06 16:09 UTC, Nanley Chery
Details

Description Nanley Chery 2016-06-02 23:19:23 UTC
Created attachment 124292 [details]
Expected result

Demo: https://github.com/SaschaWillems/Vulkan/tree/master/bloom

As of commit fb2a5ceb3264123e94a5e3f4d92cf6ec605e76e8,
anv: Emit DRAWING_RECTANGLE once at driver initialization,

the bloom demo displays rendering errors. Since the application seems to use Scissors and Viewports appropriately, there may be a bug in the Scissor/Viewport state emission code.
Comment 1 Nanley Chery 2016-06-02 23:19:50 UTC
Created attachment 124293 [details]
Actual result
Comment 2 Nanley Chery 2016-06-06 16:08:08 UTC
Once the fix which closed the following issue was pushed,
https://github.com/SaschaWillems/Vulkan/issues/70 , the rendering situation has improved. This suggests that this is synchronization-related application bug.
Comment 3 Nanley Chery 2016-06-06 16:09:44 UTC
Created attachment 124363 [details]
Actual result with synchronization fix
Comment 4 Jason Ekstrand 2016-06-06 16:11:37 UTC
(In reply to Nanley Chery from comment #2)
> Once the fix which closed the following issue was pushed,
> https://github.com/SaschaWillems/Vulkan/issues/70 , the rendering situation
> has improved. This suggests that this is synchronization-related application
> bug.

That wouldn't surprise me.  He's probably missing a pipeline barrier someplace.  We did have an issue with the CTS with this change where the DRAWING_RECTANGLE was causing a stall needed by the application and without it sum buffer flushes weren't happening as they should.  Maybe something similar is going on here.
Comment 5 Nanley Chery 2016-09-23 22:26:15 UTC
This bug no longer exists in the latest master.


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.