i965: Fix fast depth clears for surfaces with a dimension of 16384.

I hadn't bothered to set this bit because I figured it would just
paper over us getting the rectangle wrong.  But it turns out that
there is a legitimate reason to use it, so let's do so.

The alternative would be to chop up 16k clears to multiple 8k clears,
which is pointlessly painful.

Cc: "17.0" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
This commit is contained in:
Kenneth Graunke 2017-01-23 11:57:21 -08:00
parent 022e5c7e5a
commit 5106df85da
1 changed files with 12 additions and 0 deletions

View File

@ -477,6 +477,18 @@ gen8_hiz_exec(struct brw_context *brw, struct intel_mipmap_tree *mt,
break;
case BLORP_HIZ_OP_DEPTH_CLEAR:
dw1 |= GEN8_WM_HZ_DEPTH_CLEAR;
/* The "Clear Rectangle X Max" (and Y Max) fields are exclusive,
* rather than inclusive, and limited to 16383. This means that
* for a 16384x16384 render target, we would miss the last row
* or column of pixels along the edge.
*
* To work around this, we have to set the "Full Surface Depth
* and Stencil Clear" bit. We can do this in all cases because
* we always clear the full rectangle anyway. We'll need to
* change this if we ever add scissored clear support.
*/
dw1 |= GEN8_WM_HZ_FULL_SURFACE_DEPTH_CLEAR;
break;
case BLORP_HIZ_OP_NONE:
unreachable("Should not get here.");