[Intel-gfx] Mesa (master): i965: Don't write mrf assignment for pointsize output

Shuang He shuang.he at intel.com
Fri Nov 26 01:37:25 CET 2010


For you information:
This commit fixes 17 OpenGL ES 2.0 conformance cases on i965
And also another 1 commit from Ken also fixes 1 case
So We have 18 more passes on OpenGL ES 2.0 conformance on i965 these two 
days

Thanks
     --Shuang

On 2010-11-25 0:29, Kristian Høgsberg wrote:
> Module: Mesa
> Branch: master
> Commit: a889f9ee5cccee031c1090a6ef92cba894b1d77c
> URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=a889f9ee5cccee031c1090a6ef92cba894b1d77c
>
> Author: Kristian Høgsberg<krh at bitplanet.net>
> Date:   Wed Nov 24 10:12:20 2010 -0500
>
> i965: Don't write mrf assignment for pointsize output
>
> https://bugs.freedesktop.org/show_bug.cgi?id=31894
>
> ---
>
>   src/mesa/drivers/dri/i965/brw_vs_emit.c |    3 ++-
>   1 files changed, 2 insertions(+), 1 deletions(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_vs_emit.c b/src/mesa/drivers/dri/i965/brw_vs_emit.c
> index a13c3ca..89e4caf 100644
> --- a/src/mesa/drivers/dri/i965/brw_vs_emit.c
> +++ b/src/mesa/drivers/dri/i965/brw_vs_emit.c
> @@ -281,7 +281,6 @@ static void brw_vs_alloc_regs( struct brw_vs_compile *c )
>   	 else if (i == VERT_RESULT_PSIZ) {
>   	    c->regs[PROGRAM_OUTPUT][i] = brw_vec8_grf(reg, 0);
>   	    reg++;
> -	    mrf++;		/* just a placeholder?  XXX fix later stages&  remove this */
>   	 }
>   	 else {
>   	    /* Two restrictions on our compute-to-MRF here.  The
> @@ -1603,6 +1602,8 @@ static void emit_vertex_write( struct brw_vs_compile *c)
>   	 break;
>         if (!(c->prog_data.outputs_written&  BITFIELD64_BIT(i)))
>   	 continue;
> +      if (i == VERT_RESULT_PSIZ)
> +	 continue;
>
>         if (i>= VERT_RESULT_TEX0&&
>   	c->regs[PROGRAM_OUTPUT][i].file == BRW_GENERAL_REGISTER_FILE) {
>
> _______________________________________________
> mesa-commit mailing list
> mesa-commit at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-commit




More information about the Intel-gfx mailing list