[Mesa-dev] [PATCH] i965: Use tex_mocs instead of rb_mocs for GL images.
Ben Widawsky
ben at bwidawsk.net
Tue Jul 19 05:58:31 UTC 2016
On Mon, Jul 18, 2016 at 07:08:46PM -0700, Kenneth Graunke wrote:
> Fixes a 10-20% performance regression in OglCSDof caused by commit
> 5a8c89038abab0184ea72664ab390ec6ca58b4d6, which made images (in the
> image load/store sense) use BDW_MOCS_PTE instead of BDW_MOCS_WB.
>
> This seems sketchy, as the default PTE value is supposed to be
> WB LLC eLLC, which is the same as our MOCS WB setting. It's only
> supposed to change when using a surface for display, which won't
> ever happen for images. Something may be wrong in the kernel...
Which platform was the regression on? Broadwell has some weirdness if you
disable PPGTT which might cause PTE MOCS to fall back to UC. I think at least
aliasing PPGTT has been enabled since pretty early for Broadwell. If there is no
ppgtt overrides here, then I'm not sure what would be going on - it sound
sketchy to me too.
>
> Signed-off-by: Kenneth Graunke <kenneth at whitecape.org>
> Reviewed-by: Jason Ekstrand <jason at jlekstrand.net>
I haven't really followed any of the isl stuff, but it lgtm.
> ---
> src/mesa/drivers/dri/i965/brw_wm_surface_state.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
> index d896789..87f8601 100644
> --- a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
> +++ b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
> @@ -1516,7 +1516,7 @@ update_image_surface(struct brw_context *brw,
> const int surf_index = surf_offset - &brw->wm.base.surf_offset[0];
>
> brw_emit_surface_state(brw, mt, &view,
> - surface_state_infos[brw->gen].rb_mocs, false,
> + surface_state_infos[brw->gen].tex_mocs, false,
> surf_offset, surf_index,
> I915_GEM_DOMAIN_SAMPLER,
> access == GL_READ_ONLY ? 0 :
> --
> 2.9.0
>
--
Ben Widawsky, Intel Open Source Technology Center
More information about the mesa-dev
mailing list