In the updated gles3 conformance tests, this test is called ES3-CTS.gtf.GL3Tests.shadow.shadow_execution_vert.  Are you able to reproduce the failure there?

(In reply to comment #1)
> In the updated gles3 conformance tests, this test is called
> ES3-CTS.gtf.GL3Tests.shadow.shadow_execution_vert.  Are you able to
> reproduce the failure there?

Test the ES3-CTS.gtf.GL3Tests.shadow.shadow_execution_vert, the issue only can be reproduced on SNB. It works well on IVB and HSW.

(In reply to comment #2)
In the updated gles3 conformance tests, with mesa master, the issue can't be reproduced on SNB+.
> the issue only can be reproduced on SNB.
Sorry, it's my mistake. The issue can be reproduced with Mesa 10.2.4 (git-816d37e).

Run ES3-CTS.gtf.GL3Tests.shadow.shadow_execution_vert on SNB and HSW platforms.
It works well on Mesa master branch, fails on 10.2 branch.

Run it on OTC's tree, It fails on mesa master branch, works well on 10.2 branch.

Fixed on mesa master by:
commit 9548ba6e7bb8e631f6ef6236f7425725239742fe
Author: Anuj Phogat <anuj.phogat@gmail.com>
Date:   Mon Jul 21 16:58:42 2014 -0700

    mesa: Don't use memcpy() in _mesa_texstore() for float depth texture data
    
    because float depth texture data needs clamping to [0.0, 1.0]. Let the
    _mesa_texstore() fallback to slower path.
    
    Fixes Khronos GLES3 CTS tests:
    shadow_execution_vert
    shadow_execution_frag
    
    V2: Move the check to _mesa_texstore_can_use_memcpy() function.
        Add check for floating point data types.
    
    Cc: <mesa-stable@lists.freedesktop.org>
    Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>

(In reply to comment #6)
> Fixed on mesa master by:
> commit 9548ba6e7bb8e631f6ef6236f7425725239742fe
> Author: Anuj Phogat <anuj.phogat@gmail.com>
> Date:   Mon Jul 21 16:58:42 2014 -0700
> 
>     mesa: Don't use memcpy() in _mesa_texstore() for float depth texture data
>     
>     because float depth texture data needs clamping to [0.0, 1.0]. Let the
>     _mesa_texstore() fallback to slower path.
>     
>     Fixes Khronos GLES3 CTS tests:
>     shadow_execution_vert
>     shadow_execution_frag
>     
>     V2: Move the check to _mesa_texstore_can_use_memcpy() function.
>         Add check for floating point data types.
>     
>     Cc: <mesa-stable@lists.freedesktop.org>
>     Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
>     Reviewed-by: Marek Olšák <marek.olsak@amd.com>


Run on OTC'tree and mesa master commit 9276ef6f41626307, it still fail.

Created attachment 104467 [details]
output

Someone has missed something.

The original report says a commit that claimed to fix two tests in fact broke them. Then comment six just restates the claim without seeming to acknowledge the report says it broke the tests.

This bug (description and comments) is certainly hard to read/understand. I've run the test on IVB laptop against current Mesa master (58a5409) like this:

./glcts --deqp-case=ES3-CTS.gtf.GL3Tests.shadow.shadow_execution_vert --deqp-gl-config-id=14

and got:

DONE!

Test run totals:
  Passed:        1/1 (100.00%)
  Failed:        0/1 (0.00%)
  Not supported: 0/1 (0.00%)
  Warnings:      0/1 (0.00%)

as a result. Does this case fail on some particular platform (like stated in some of the comments)?

(In reply to Tapani Pälli from comment #10)
> This bug (description and comments) is certainly hard to read/understand.
> I've run the test on IVB laptop against current Mesa master (58a5409) like
> this:
> 
> ./glcts --deqp-case=ES3-CTS.gtf.GL3Tests.shadow.shadow_execution_vert
> --deqp-gl-config-id=14
> 
> and got:
> 
> DONE!
> 
> Test run totals:
>   Passed:        1/1 (100.00%)
>   Failed:        0/1 (0.00%)
>   Not supported: 0/1 (0.00%)
>   Warnings:      0/1 (0.00%)
> 
> as a result. Does this case fail on some particular platform (like stated in
> some of the comments)?

Test on latest Mesa Master branch.
On OTC's tree, Run ./GTF -width=64 -height=64 -run=GL3Tests/shadow/shadow_execution_frag.tes ,it still fail.
Skipping ConfigID: 82 Sample Buffers: 1 Samples: 4
Skipping ConfigID: 84 Sample Buffers: 1 Samples: 8
Skipping ConfigID: 86 Sample Buffers: 1 Samples: 4
Skipping ConfigID: 88 Sample Buffers: 1 Samples: 8
Conformance FAILED 6 of 6 tests.
Test duration: 3s

On gles3-new tree, run ./glcts --deqp-case=ES3-CTS.gtf.GL3Tests.shadow.shadow_execution_vert --deqp-gl-config-id=14 ,it works well.
Test run totals:
  Passed:        1/1 (100.00%)
  Failed:        0/1 (0.00%)
  Not supported: 0/1 (0.00%)
  Warnings:      0/1 (0.00%)

What is the status of this bug? Can we close it?

Check the result again.  
OTC's tree has one case issue, it still fail, but driver is OK.
Ogles3-new tree is works well.  Close it.

Verified.Fixed.