[Bug 109113] KHR-GL43.compute_shader.resource-texture fails with MCS fast clears disabled

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Thu Dec 20 08:01:02 UTC 2018


https://bugs.freedesktop.org/show_bug.cgi?id=109113

            Bug ID: 109113
           Summary: KHR-GL43.compute_shader.resource-texture fails with
                    MCS fast clears disabled
           Product: Mesa
           Version: git
          Hardware: Other
                OS: All
            Status: NEW
          Severity: normal
          Priority: medium
         Component: Drivers/DRI/i965
          Assignee: intel-3d-bugs at lists.freedesktop.org
          Reporter: kenneth at whitecape.org
        QA Contact: intel-3d-bugs at lists.freedesktop.org
                CC: ghugesss at gmail.com, jason at jlekstrand.net,
                    nanleychery at gmail.com, topi.pohjolainen at intel.com

Created attachment 142859
  --> https://bugs.freedesktop.org/attachment.cgi?id=142859&action=edit
Patch to narrow the CTS test to a smaller reproducer

I believe I've found a bug with MCS and slow clears.

If I disable fast clears (hack brw_blorp.c to add "can_fast_clear = false"),
then KHR-GL43.compute_shader.resource-texture fails on my Kabylake system.  The
test clears a 4x MSAA 4x16 R32G32B32A32_FLOAT texture, which is using the CMS
layout.  Disabling CMS/MCS also fixes it.  (There are actually two, one is a
single slice, the other is a 2-slice array texture.  Both are broken.)  It then
samples the textures with a compute shader.

Fast clears ought to be optional - I think they may be masking a bug.

The attached patch to the CTS test narrows it down to a much smaller test case.

This is fairly outside of my expertise, so I'd appreciate it if some MCS
experts could take a look (without ruining your holiday plans)...thanks!

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20181220/7268e82d/attachment-0001.html>


More information about the intel-3d-bugs mailing list