[Bug 97477] i915g: gl_FragCoord is always (0.0, max_y)

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Thu Aug 25 05:16:17 UTC 2016


https://bugs.freedesktop.org/show_bug.cgi?id=97477

            Bug ID: 97477
           Summary: i915g: gl_FragCoord is always (0.0, max_y)
           Product: Mesa
           Version: git
          Hardware: Other
                OS: All
            Status: NEW
          Severity: normal
          Priority: medium
         Component: Drivers/Gallium/i915g
          Assignee: dri-devel at lists.freedesktop.org
          Reporter: nicholasbishop at gmail.com
        QA Contact: dri-devel at lists.freedesktop.org

Created attachment 126029
  --> https://bugs.freedesktop.org/attachment.cgi?id=126029&action=edit
Patch: Handle I915_SEMANTIC_POS when setting texCoords for gl_FragCoord

On Intel Pineview M hardware, the i915 gallium driver doesn't output the
correct gl_FragCoord. It seems to always have an X coord of 0.0 and a Y coord
of the window's height in pixels, e.g. 600.0f or such.

I believe this is a regression caused in part by commit
afa035031ff9e0c07a2297d864e46c76f7bfff58
(https://lists.freedesktop.org/archives/mesa-commit/2016-January/061181.html)

The old behavior used output zero, while the new behavior uses actual zeroes.
In the case of gl_FragCoord output zero happened to be the correct one.

I've attached a patch that seems to fix the issue for me. I ran the piglit
"quick" tests and observed 17 new tests passing, and one regression. I suspect
the regression is unimportant since it involves derivatives that aren't
implemented for this driver anyway:

"fs-dfdy-accuracy" test failed ... in i915_fpc_translate.c:
case TGSI_OPCODE_DDY:
      /* XXX We just output 0 here */
      debug_printf("Punting DDX/DDX\n");

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/dri-devel/attachments/20160825/2c39b4bd/attachment-0001.html>


More information about the dri-devel mailing list