i965: do not fallback to linear tiling for stencil surfaces

We were skipping this fallback for depth, but not for stencil
which the hardware always requires to be W-tiled.

Also, make the checks for whether we need to apply retiling
strategies based on usage instead of tiling flags, which is
safer and more explicit.

This fixes a regression in a CTS test introduced with commit
4ea63fab77 that started applying re-tiling stencil surfaces
in certain scenarios.

v2: discard retiling based on usage fields instead of tiling
    flags. This is safer and more explicit.

v3: Add a comment indicating that texturing of stencil in gen7
    requires an Y-tiled copy (Topi).

Fixes:
KHR-GL45.direct_state_access.renderbuffers_storage

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
This commit is contained in:
Iago Toral Quiroga 2017-09-07 12:56:27 +02:00
parent 806ae6a648
commit ab6f874439

View file

@ -570,11 +570,14 @@ make_surface(struct brw_context *brw, GLenum target, mesa_format format,
if (!isl_surf_init_s(&brw->isl_dev, &mt->surf, &init_info))
goto fail;
/* In case caller doesn't specifically request Y-tiling (needed
* unconditionally for depth), check for corner cases needing special
* treatment.
/* Depth surfaces are always Y-tiled and stencil is always W-tiled, although
* on gen7 platforms we also need to create Y-tiled copies of stencil for
* texturing since the hardware can't sample from W-tiled surfaces. For
* everything else, check for corner cases needing special treatment.
*/
if (tiling_flags & ~ISL_TILING_Y0_BIT) {
bool is_depth_stencil =
mt->surf.usage & (ISL_SURF_USAGE_STENCIL_BIT | ISL_SURF_USAGE_DEPTH_BIT);
if (!is_depth_stencil) {
if (need_to_retile_as_linear(brw, mt->surf.row_pitch,
mt->surf.tiling, mt->surf.samples)) {
init_info.tiling_flags = 1u << ISL_TILING_LINEAR;