Re: [Intel-gfx] [Mesa-dev] [PATCH 1/9] intel gen4-5: fix the vue view in the fs.

2012-07-27 Thread Olivier Galibert
On Thu, Jul 26, 2012 at 10:18:01AM -0700, Eric Anholt wrote:
 Olivier Galibert galib...@pobox.com writes:
 
  In some cases the fragment shader view of the vue registers was out of
  sync with the builder.  This fixes it.
 
 s/builder/SF outputs/ ?
 
 I'd love to see the pre-gen6 code get rearranged so the FS walked the
 bitfield of FS inputs from SF and chose the urb offset for each.  But
 this does look like the minimal fix.

In other words, an explicit linking pass?  That could be useful with
geometry shaders, too.

  OG.
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx


Re: [Intel-gfx] [Mesa-dev] [PATCH 1/9] intel gen4-5: fix the vue view in the fs.

2012-07-26 Thread Eric Anholt
Olivier Galibert galib...@pobox.com writes:

 In some cases the fragment shader view of the vue registers was out of
 sync with the builder.  This fixes it.

s/builder/SF outputs/ ?

I'd love to see the pre-gen6 code get rearranged so the FS walked the
bitfield of FS inputs from SF and chose the urb offset for each.  But
this does look like the minimal fix.

Reviewed-by: Eric Anholt e...@anholt.net


pgp5F8dGu9BYv.pgp
Description: PGP signature
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx