AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

intel: Don't use HW clear color conversion on gfx11

Browse Source 
The hardware's clear color conversion feature unfortunately requires
invalidating the texture cache for every fast clear. To avoid the
performance penalty that comes with the invalidation, avoid using the
hardware feature and write out the converted clear color pixel
ourselves.

When testing a patch which moves a state cache invalidate to occur after
fast clears instead of before, this prevents the following failures on
icl/zink:

* piglit.fast_color_clear.fcc-read-after-clear sample tex
* piglit.spec.arb_clear_texture.arb_clear_texture-cube

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30646>
This commit is contained in:
Nanley Chery
2024-06-24 16:09:57 -04:00
committed by Marge Bot
parent dfcd93d12f
commit b404ca0eb0
2 changed files with 13 additions and 48 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/intel/blorp/blorp_genX_exec_brw.h
+13 -28
View File
@@ -1513,36 +1513,21 @@ blorp_update_clear_color(UNUSED struct blorp_batch *batch,
const struct blorp_surface_info *info)
{
assert(info->clear_color_addr.buffer != NULL);
uint32_t pixel[4];
isl_color_value_pack(&info->clear_color, info->surf.format, pixel);
#if GFX_VER == 11
/* 2 QWORDS */
const unsigned inlinedata_dw = 2 * 2;
const unsigned num_dwords = GENX(MI_ATOMIC_length) + inlinedata_dw;
struct blorp_address clear_addr = info->clear_color_addr;
uint32_t *dw = blorp_emitn(batch, GENX(MI_ATOMIC), num_dwords,
.DataSize = MI_ATOMIC_QWORD,
.ATOMICOPCODE = MI_ATOMIC_OP_MOVE8B,
.InlineData = true,
.MemoryAddress = clear_addr);
/* dw starts at dword 1, but we need to fill dwords 3 and 5 */
uint32_t *dw = blorp_emitn(batch, GENX(MI_STORE_DATA_IMM), 3 + 6,
.StoreQword = true,
.Address = info->clear_color_addr);
/* dw starts at dword 1 */
dw[2] = info->clear_color.u32[0];
dw[3] = 0;
dw[4] = info->clear_color.u32[1];
dw[5] = 0;
clear_addr.offset += 8;
dw = blorp_emitn(batch, GENX(MI_ATOMIC), num_dwords,
.DataSize = MI_ATOMIC_QWORD,
.ATOMICOPCODE = MI_ATOMIC_OP_MOVE8B,
.CSSTALL = true,
.ReturnDataControl = true,
.InlineData = true,
.MemoryAddress = clear_addr);
/* dw starts at dword 1, but we need to fill dwords 3 and 5 */
dw[2] = info->clear_color.u32[2];
dw[3] = 0;
dw[4] = info->clear_color.u32[3];
dw[5] = 0;
dw[3] = info->clear_color.u32[1];
dw[4] = info->clear_color.u32[2];
dw[5] = info->clear_color.u32[3];
dw[6] = pixel[0];
dw[7] = pixel[1];
#else