[Mesa-dev] [RFC PATCH 03/16] i965/gen6: Align height to 2 with MSAA for certain surface heights
Pohjolainen, Topi
topi.pohjolainen at intel.com
Sun Jun 1 22:49:38 PDT 2014
On Thu, May 29, 2014 at 01:53:42PM -0700, Jordan Justen wrote:
> TODO: Find a good documented reason for this change.
>
> This could be related to PRM Volume 1 Part 1: Graphics Core, Section
> 7.18.3.7.1 (Surface Arrays For all surfaces other than separate
> stencil buffer):
>
> "[DevSNB] Errata: Sampler MSAA Qpitch will be 4 greater than the
> value calculated in the equation above , for every other odd Surface
> Height starting from 1 i.e. 1,5,9,13"
In gen6 msaa is always interleaved and there is only one option - four sample
per pixel. In the memory samples are organised for a 2x2 pixel span by
grouping samples for one index into a 2x2 block:
0 0 1 1
0 0 1 1
2 2 3 3
2 2 3 3
So I would expect alignment factor of four instead of two.
>
> Signed-off-by: Jordan Justen <jordan.l.justen at intel.com>
> ---
> src/mesa/drivers/dri/i965/gen6_surface_state.c | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/src/mesa/drivers/dri/i965/gen6_surface_state.c b/src/mesa/drivers/dri/i965/gen6_surface_state.c
> index 728488a..715111f 100644
> --- a/src/mesa/drivers/dri/i965/gen6_surface_state.c
> +++ b/src/mesa/drivers/dri/i965/gen6_surface_state.c
> @@ -101,8 +101,12 @@ gen6_update_renderbuffer_surface(struct brw_context *brw,
> /* reloc */
> surf[1] = mt->bo->offset64;
>
> + int height0 = irb->mt->logical_height0;
> + if (brw->gen == 6 && irb->mt->num_samples > 1 && (height0 % 4) == 1)
> + height0 = ALIGN(height0, 2);
> +
> surf[2] = ((irb->mt->logical_width0 - 1) << BRW_SURFACE_WIDTH_SHIFT |
> - (irb->mt->logical_height0 - 1) << BRW_SURFACE_HEIGHT_SHIFT |
> + (height0 - 1) << BRW_SURFACE_HEIGHT_SHIFT |
> ((irb->mt_level - irb->mt->first_level)) << BRW_SURFACE_LOD_SHIFT);
>
> surf[3] = (brw_get_surface_tiling_bits(mt->tiling) |
> --
> 2.0.0.rc4
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
More information about the mesa-dev
mailing list