Bug 105232 - piglit regressions and crashes from gbm alpha masks
Summary: piglit regressions and crashes from gbm alpha masks
Status: RESOLVED NOTABUG
Alias: None
Product: Mesa
Classification: Unclassified
Component: Mesa core (show other bugs)
Version: git
Hardware: Other All
: medium normal
Assignee: Eric Anholt
QA Contact: mesa-dev
URL:
Whiteboard:
Keywords: bisected, regression
Depends on:
Blocks:
 
Reported: 2018-02-24 08:51 UTC by Mark Janes
Modified: 2018-04-23 13:04 UTC (History)
3 users (show)

See Also:
i915 platform:
i915 features:


Attachments

Description Mark Janes 2018-02-24 08:51:56 UTC
Most i965 platforms fail 3 tests after 
880573e7370b866c94513f59ad0bb0b10660643f
Author:     Eric Anholt <eric@anholt.net>

gbm: Fix the alpha masks in the GBM format table.

Once GBM started looking at the values of the alpha masks, ARGB/ABGR
wouldn't match any more because we had both A and R in the low bits.

Fixes: 2ed344645d65 ("gbm/dri: Add RGBA masks to GBM format table")
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Daniel Stone <daniels@collabora.com>

    piglit.fast_color_clear.fcc-front-buffer-distraction.kblm32
    piglit.spec.oes_packed_depth_stencil.depth_stencil texture gles1.kblm32
    piglit.spec.!opengl 1_0.gl-1_0-swapbuffers-behavior.kblm32

Sample output:
/tmp/build_root/m32/lib/piglit/bin/oes_packed_depth_stencil-depth-stencil-texture_gles1 -auto -fbo
Testing glTexImage with GL_DEPTH_STENCIL...
Unexpected GL error: GL_INVALID_OPERATION 0x502
(Error at ../tests/spec/ext_packed_depth_stencil/depth-stencil-texture.c:308)

i915 platforms have more regressions (34 on g33).
Comment 1 Ilia Mirkin 2018-02-24 17:42:49 UTC
Theory: The tests didn't run prior to this commit because of the format matching fail, since the original commit that broke argb8 formats, 2ed344645d65. They have now started running again and are as broken now as they were prior to that commit.

Mark, does that theory hold water?
Comment 2 Daniel Stone 2018-03-16 16:33:14 UTC
(In reply to Ilia Mirkin from comment #1)
> Theory: The tests didn't run prior to this commit because of the format
> matching fail, since the original commit that broke argb8 formats,
> 2ed344645d65. They have now started running again and are as broken now as
> they were prior to that commit.
> 
> Mark, does that theory hold water?

Any idea?


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.