OpenHarmony/third_party_mesa3d
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

freedreno/a6xx: Remove tex fb_read state

Browse Source 
Now that fb_read is using a bindless descriptor slot, we can remove the
dependency of the TEX state on the PROG state.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20687>
This commit is contained in:
Rob Clark
2023-01-01 15:11:20 -08:00
committed by Marge Bot
parent f6b6400254
commit b17ad0657f
4 changed files with 18 additions and 75 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
src/gallium/drivers/freedreno/a6xx/fd6_context.c
View File

@@ -167,7 +167,6 @@ setup_state_map(struct fd_context *ctx)
fd_context_add_map(ctx, FD_DIRTY_BLEND | FD_DIRTY_SAMPLE_MASK,
BIT(FD6_GROUP_BLEND));
fd_context_add_map(ctx, FD_DIRTY_BLEND_COLOR, BIT(FD6_GROUP_BLEND_COLOR));
fd_context_add_map(ctx, FD_DIRTY_PROG, BIT(FD6_GROUP_FS_TEX));
fd_context_add_map(ctx, FD_DIRTY_PROG | FD_DIRTY_CONST,
BIT(FD6_GROUP_CONST));
fd_context_add_map(ctx, FD_DIRTY_STREAMOUT, BIT(FD6_GROUP_SO));

87
src/gallium/drivers/freedreno/a6xx/fd6_emit.c
View File

@@ -50,22 +50,9 @@
#include "fd6_texture.h"
#include "fd6_zsa.h"
static void
fd6_emit_fb_tex(struct fd_ringbuffer *state, struct fd_context *ctx) assert_dt
{
/* This is patched with the appropriate descriptor for GMEM or
* sysmem rendering path in fd6_gmem
*/
OUT_RINGP(state, 0, &ctx->batch->fb_read_patches);
for (unsigned i = 1; i < FDL6_TEX_CONST_DWORDS; i++)
OUT_RING(state, 0);
}
void
fd6_emit_textures(struct fd_context *ctx, struct fd_ringbuffer *ring,
enum pipe_shader_type type, struct fd_texture_stateobj *tex,
/* can be NULL if no image/SSBO/fb state to merge in: */
const struct ir3_shader_variant *v)
enum pipe_shader_type type, struct fd_texture_stateobj *tex)
{
unsigned opcode, tex_samp_reg, tex_const_reg, tex_count_reg;
enum a6xx_state_block sb;
@@ -146,14 +133,11 @@ fd6_emit_textures(struct fd_context *ctx, struct fd_ringbuffer *ring,
fd_ringbuffer_del(state);
}
unsigned num_merged_textures = tex->num_textures;
unsigned num_textures = tex->num_textures;
if (v && v->fb_read)
num_merged_textures++;
if (num_merged_textures > 0) {
if (num_textures > 0) {
struct fd_ringbuffer *state =
fd_ringbuffer_new_object(ctx->pipe, num_merged_textures * 16 * 4);
fd_ringbuffer_new_object(ctx->pipe, num_textures * 16 * 4);
for (unsigned i = 0; i < num_textures; i++) {
const struct fd6_pipe_sampler_view *view;
@@ -200,16 +184,13 @@ fd6_emit_textures(struct fd_context *ctx, struct fd_ringbuffer *ring,
OUT_RING(state, view->descriptor[15]);
}
if (v && v->fb_read)
fd6_emit_fb_tex(state, ctx);
/* emit texture state: */
OUT_PKT7(ring, opcode, 3);
OUT_RING(ring, CP_LOAD_STATE6_0_DST_OFF(0) |
CP_LOAD_STATE6_0_STATE_TYPE(ST6_CONSTANTS) |
CP_LOAD_STATE6_0_STATE_SRC(SS6_INDIRECT) |
CP_LOAD_STATE6_0_STATE_BLOCK(sb) |
CP_LOAD_STATE6_0_NUM_UNIT(num_merged_textures));
CP_LOAD_STATE6_0_NUM_UNIT(num_textures));
OUT_RB(ring, state); /* SRC_ADDR_LO/HI */
OUT_PKT4(ring, tex_const_reg, 2);
@@ -219,7 +200,7 @@ fd6_emit_textures(struct fd_context *ctx, struct fd_ringbuffer *ring,
}
OUT_PKT4(ring, tex_count_reg, 1);
OUT_RING(ring, num_merged_textures);
OUT_RING(ring, num_textures);
}
/* Emits combined texture state, which also includes any Image/SSBO
@@ -234,8 +215,7 @@ fd6_emit_textures(struct fd_context *ctx, struct fd_ringbuffer *ring,
*/
static void
fd6_emit_combined_textures(struct fd6_emit *emit,
enum pipe_shader_type type,
const struct ir3_shader_variant *v) assert_dt
enum pipe_shader_type type) assert_dt
{
struct fd_context *ctx = emit->ctx;
@@ -249,38 +229,13 @@ fd6_emit_combined_textures(struct fd6_emit *emit,
assert((type < ARRAY_SIZE(s)) && s[type]);
if (!v->fb_read) {
/* in the fast-path, when we don't have to mix in any image/SSBO
* related texture state, we can just lookup the stateobj and
* re-emit that:
*
* Also, framebuffer-read is a slow-path because an extra
* texture needs to be inserted.
*/
if (ctx->tex[type].num_textures > 0) {
struct fd6_texture_state *tex =
if (ctx->tex[type].num_textures > 0) {
struct fd6_texture_state *tex =
fd6_texture_state(ctx, type, &ctx->tex[type]);
fd6_state_add_group(&emit->state, tex->stateobj, s[type]);
fd6_state_add_group(&emit->state, tex->stateobj, s[type]);
fd6_texture_state_reference(&tex, NULL);
}
} else {
/* In the slow-path, create a one-shot texture state object
* if either TEX|PROG|SSBO|IMAGE state is dirty:
*/
if ((ctx->dirty_shader[type] &
(FD_DIRTY_SHADER_TEX | FD_DIRTY_SHADER_PROG | FD_DIRTY_SHADER_IMAGE |
FD_DIRTY_SHADER_SSBO)) ||
v->fb_read) {
struct fd_texture_stateobj *tex = &ctx->tex[type];
struct fd_ringbuffer *stateobj = fd_submit_new_ringbuffer(
ctx->batch->submit, 0x1000, FD_RINGBUFFER_STREAMING);
fd6_emit_textures(ctx, stateobj, type, tex, v);
fd6_state_take_group(&emit->state, stateobj, s[type]);
}
fd6_texture_state_reference(&tex, NULL);
}
}
@@ -712,10 +667,6 @@ fd6_emit_3d_state(struct fd_ringbuffer *ring, struct fd6_emit *emit)
struct fd_context *ctx = emit->ctx;
struct pipe_framebuffer_state *pfb = &ctx->batch->framebuffer;
const struct fd6_program_state *prog = fd6_emit_get_prog(emit);
const struct ir3_shader_variant *vs = emit->vs;
const struct ir3_shader_variant *hs = emit->hs;
const struct ir3_shader_variant *ds = emit->ds;
const struct ir3_shader_variant *gs = emit->gs;
const struct ir3_shader_variant *fs = emit->fs;
emit_marker6(ring, 5);
@@ -820,25 +771,19 @@ fd6_emit_3d_state(struct fd_ringbuffer *ring, struct fd6_emit *emit)
fd6_state_take_group(&emit->state, state, FD6_GROUP_PRIMITIVE_PARAMS);
break;
case FD6_GROUP_VS_TEX:
fd6_emit_combined_textures(emit, PIPE_SHADER_VERTEX, vs);
fd6_emit_combined_textures(emit, PIPE_SHADER_VERTEX);
break;
case FD6_GROUP_HS_TEX:
if (hs) {
fd6_emit_combined_textures(emit, PIPE_SHADER_TESS_CTRL, hs);
}
fd6_emit_combined_textures(emit, PIPE_SHADER_TESS_CTRL);
break;
case FD6_GROUP_DS_TEX:
if (ds) {
fd6_emit_combined_textures(emit, PIPE_SHADER_TESS_EVAL, ds);
}
fd6_emit_combined_textures(emit, PIPE_SHADER_TESS_EVAL);
break;
case FD6_GROUP_GS_TEX:
if (gs) {
fd6_emit_combined_textures(emit, PIPE_SHADER_GEOMETRY, gs);
}
fd6_emit_combined_textures(emit, PIPE_SHADER_GEOMETRY);
break;
case FD6_GROUP_FS_TEX:
fd6_emit_combined_textures(emit, PIPE_SHADER_FRAGMENT, fs);
fd6_emit_combined_textures(emit, PIPE_SHADER_FRAGMENT);
break;
case FD6_GROUP_SO:
fd6_emit_streamout(ring, emit);
@@ -865,7 +810,7 @@ fd6_emit_cs_state(struct fd_context *ctx, struct fd_ringbuffer *ring,
FD_DIRTY_SHADER_IMAGE | FD_DIRTY_SHADER_SSBO)) {
struct fd_texture_stateobj *tex = &ctx->tex[PIPE_SHADER_COMPUTE];
fd6_emit_textures(ctx, ring, PIPE_SHADER_COMPUTE, tex, cp);
fd6_emit_textures(ctx, ring, PIPE_SHADER_COMPUTE, tex);
OUT_PKT4(ring, REG_A6XX_SP_VS_TEX_COUNT, 1);
OUT_RING(ring, 0);

3
src/gallium/drivers/freedreno/a6xx/fd6_emit.h
View File

@@ -323,8 +323,7 @@ fd6_gl2spacing(enum gl_tess_spacing spacing)
void fd6_emit_textures(struct fd_context *ctx, struct fd_ringbuffer *ring,
enum pipe_shader_type type,
struct fd_texture_stateobj *tex,
const struct ir3_shader_variant *v) assert_dt;
struct fd_texture_stateobj *tex) assert_dt;
void fd6_emit_3d_state(struct fd_ringbuffer *ring,
struct fd6_emit *emit) assert_dt;

2
src/gallium/drivers/freedreno/a6xx/fd6_texture.c
View File

@@ -562,7 +562,7 @@ fd6_texture_state(struct fd_context *ctx, enum pipe_shader_type type,
state->key = key;
state->stateobj = fd_ringbuffer_new_object(ctx->pipe, 32 * 4);
fd6_emit_textures(ctx, state->stateobj, type, tex, NULL);
fd6_emit_textures(ctx, state->stateobj, type, tex);
/* NOTE: uses copy of key in state obj, because pointer passed by caller
* is probably on the stack