[Mesa-dev] [PATCH] i965: Drop mark_surface_used mechanism.
Kenneth Graunke
kenneth at whitecape.org
Sun Dec 30 06:01:34 UTC 2018
The original idea was that the backend compiler could eliminate
surfaces, so we would have it mark which ones are actually used,
then shrink the binding table accordingly. Unfortunately, it's a
pretty blunt mechanism - it can only prune things from the end,
not the middle - since we decide the layout before we even start
the backend compiler, and only limit the size. It also basically
gives up if it sees indirect array access.
Besides, we do the vast majority of our surface elimination in NIR
anyway, not the backend - and I don't see that trend changing any
time soon. Vulkan abandoned this plan a long time ago, and I don't
use it in Iris, but it's still been kicking around in i965.
---
src/intel/compiler/brw_compiler.h | 13 --------
src/intel/compiler/brw_fs.cpp | 4 ---
src/intel/compiler/brw_fs_generator.cpp | 11 -------
src/intel/compiler/brw_fs_nir.cpp | 32 -------------------
src/intel/compiler/brw_vec4_generator.cpp | 9 ------
src/intel/compiler/brw_vec4_nir.cpp | 29 -----------------
src/intel/compiler/brw_vec4_visitor.cpp | 2 --
.../drivers/dri/i965/brw_nir_uniforms.cpp | 4 ---
src/mesa/drivers/dri/i965/brw_program.c | 5 ++-
src/mesa/drivers/dri/i965/brw_wm.c | 3 ++
10 files changed, 7 insertions(+), 105 deletions(-)
I thought this would let me eliminate nir_tex_instr::texture_array_size
too, but apparently we use that for the used_by_txf sRGB shenanigans.
Ultimately, I'd like to start using nir_lower_samplers_as_deref instead
of nir_lower_samplers, and this is one step towards that.
diff --git a/src/intel/compiler/brw_compiler.h b/src/intel/compiler/brw_compiler.h
index e4f4d83c8e8..61a4528d372 100644
--- a/src/intel/compiler/brw_compiler.h
+++ b/src/intel/compiler/brw_compiler.h
@@ -643,19 +643,6 @@ brw_stage_prog_data_add_params(struct brw_stage_prog_data *prog_data,
return prog_data->param + old_nr_params;
}
-static inline void
-brw_mark_surface_used(struct brw_stage_prog_data *prog_data,
- unsigned surf_index)
-{
- /* A binding table index is 8 bits and the top 3 values are reserved for
- * special things (stateless and SLM).
- */
- assert(surf_index <= 252);
-
- prog_data->binding_table.size_bytes =
- MAX2(prog_data->binding_table.size_bytes, (surf_index + 1) * 4);
-}
-
enum brw_barycentric_mode {
BRW_BARYCENTRIC_PERSPECTIVE_PIXEL = 0,
BRW_BARYCENTRIC_PERSPECTIVE_CENTROID = 1,
diff --git a/src/intel/compiler/brw_fs.cpp b/src/intel/compiler/brw_fs.cpp
index 590a9b32a8e..d5f70e6a2f1 100644
--- a/src/intel/compiler/brw_fs.cpp
+++ b/src/intel/compiler/brw_fs.cpp
@@ -2376,8 +2376,6 @@ fs_visitor::lower_constant_loads()
inst->src[i].nr = dst.nr;
inst->src[i].offset = (base & (block_sz - 1)) +
inst->src[i].offset % 4;
-
- brw_mark_surface_used(prog_data, index);
}
if (inst->opcode == SHADER_OPCODE_MOV_INDIRECT &&
@@ -2391,8 +2389,6 @@ fs_visitor::lower_constant_loads()
inst->src[1],
pull_index * 4);
inst->remove(block);
-
- brw_mark_surface_used(prog_data, index);
}
}
invalidate_live_intervals();
diff --git a/src/intel/compiler/brw_fs_generator.cpp b/src/intel/compiler/brw_fs_generator.cpp
index 08dd83dded7..97a1760528c 100644
--- a/src/intel/compiler/brw_fs_generator.cpp
+++ b/src/intel/compiler/brw_fs_generator.cpp
@@ -315,8 +315,6 @@ fs_generator::fire_fb_write(fs_inst *inst,
if (devinfo->gen >= 6)
brw_inst_set_rt_slot_group(devinfo, insn, inst->group / 16);
-
- brw_mark_surface_used(&prog_data->base, surf_index);
}
void
@@ -373,8 +371,6 @@ fs_generator::generate_fb_read(fs_inst *inst, struct brw_reg dst,
gen9_fb_READ(p, dst, payload, surf_index,
inst->header_size, inst->size_written / REG_SIZE,
prog_data->persample_dispatch);
-
- brw_mark_surface_used(&prog_data->base, surf_index);
}
void
@@ -872,8 +868,6 @@ fs_generator::generate_get_buffer_size(fs_inst *inst,
inst->header_size > 0,
simd_mode,
BRW_SAMPLER_RETURN_FORMAT_SINT32);
-
- brw_mark_surface_used(prog_data, surf_index.ud);
}
void
@@ -1158,8 +1152,6 @@ fs_generator::generate_tex(fs_inst *inst, struct brw_reg dst, struct brw_reg src
inst->header_size != 0,
simd_mode,
return_format);
-
- brw_mark_surface_used(prog_data, surface + base_binding_table_index);
} else {
/* Non-const sampler index */
@@ -1752,9 +1744,6 @@ fs_generator::generate_shader_time_add(fs_inst *,
brw_shader_time_add(p, payload,
prog_data->binding_table.shader_time_start);
brw_pop_insn_state(p);
-
- brw_mark_surface_used(prog_data,
- prog_data->binding_table.shader_time_start);
}
void
diff --git a/src/intel/compiler/brw_fs_nir.cpp b/src/intel/compiler/brw_fs_nir.cpp
index 92ec85a27cc..b62dfe0f120 100644
--- a/src/intel/compiler/brw_fs_nir.cpp
+++ b/src/intel/compiler/brw_fs_nir.cpp
@@ -2939,10 +2939,6 @@ fs_visitor::emit_non_coherent_fb_read(const fs_builder &bld, const fs_reg &dst,
wm_prog_data->binding_table.render_target_read_start -
wm_prog_data->base.binding_table.texture_start;
- brw_mark_surface_used(
- bld.shader->stage_prog_data,
- wm_prog_data->binding_table.render_target_read_start + target);
-
/* Calculate the fragment coordinates. */
const fs_reg coords = bld.vgrf(BRW_REGISTER_TYPE_UD, 3);
bld.MOV(offset(coords, bld, 0), pixel_x);
@@ -3424,7 +3420,6 @@ fs_visitor::nir_emit_cs_intrinsic(const fs_builder &bld,
cs_prog_data->uses_num_work_groups = true;
fs_reg surf_index = brw_imm_ud(surface);
- brw_mark_surface_used(prog_data, surface);
/* Read the 3 GLuint components of gl_NumWorkGroups */
for (unsigned i = 0; i < 3; i++) {
@@ -3644,18 +3639,10 @@ fs_visitor::get_nir_ssbo_intrinsic_index(const brw::fs_builder &bld,
unsigned index = stage_prog_data->binding_table.ssbo_start +
nir_src_as_uint(instr->src[src]);
surf_index = brw_imm_ud(index);
- brw_mark_surface_used(prog_data, index);
} else {
surf_index = vgrf(glsl_type::uint_type);
bld.ADD(surf_index, get_nir_src(instr->src[src]),
brw_imm_ud(stage_prog_data->binding_table.ssbo_start));
-
- /* Assume this may touch any UBO. It would be nice to provide
- * a tighter bound, but the array information is already lowered away.
- */
- brw_mark_surface_used(prog_data,
- stage_prog_data->binding_table.ssbo_start +
- nir->info.num_ssbos - 1);
}
return surf_index;
@@ -3941,7 +3928,6 @@ fs_visitor::nir_emit_intrinsic(const fs_builder &bld, nir_intrinsic_instr *instr
const unsigned index = stage_prog_data->binding_table.ubo_start +
nir_src_as_uint(instr->src[0]);
surf_index = brw_imm_ud(index);
- brw_mark_surface_used(prog_data, index);
} else {
/* The block index is not a constant. Evaluate the index expression
* per-channel and add the base UBO index; we have to select a value
@@ -3951,13 +3937,6 @@ fs_visitor::nir_emit_intrinsic(const fs_builder &bld, nir_intrinsic_instr *instr
bld.ADD(surf_index, get_nir_src(instr->src[0]),
brw_imm_ud(stage_prog_data->binding_table.ubo_start));
surf_index = bld.emit_uniformize(surf_index);
-
- /* Assume this may touch any UBO. It would be nice to provide
- * a tighter bound, but the array information is already lowered away.
- */
- brw_mark_surface_used(prog_data,
- stage_prog_data->binding_table.ubo_start +
- nir->info.num_ubos - 1);
}
if (!nir_src_is_const(instr->src[1])) {
@@ -4214,8 +4193,6 @@ fs_visitor::nir_emit_intrinsic(const fs_builder &bld, nir_intrinsic_instr *instr
ubld.ADD(buffer_size, size_aligned4, negate(size_padding));
bld.MOV(retype(dest, ret_payload.type), component(buffer_size, 0));
-
- brw_mark_surface_used(prog_data, index);
break;
}
@@ -4826,15 +4803,6 @@ fs_visitor::nir_emit_texture(const fs_builder &bld, nir_tex_instr *instr)
unreachable("should be lowered");
case nir_tex_src_texture_offset: {
- /* Figure out the highest possible texture index and mark it as used */
- uint32_t max_used = texture + instr->texture_array_size - 1;
- if (instr->op == nir_texop_tg4 && devinfo->gen < 8) {
- max_used += stage_prog_data->binding_table.gather_texture_start;
- } else {
- max_used += stage_prog_data->binding_table.texture_start;
- }
- brw_mark_surface_used(prog_data, max_used);
-
/* Emit code to evaluate the actual indexing expression */
fs_reg tmp = vgrf(glsl_type::uint_type);
bld.ADD(tmp, src, brw_imm_ud(texture));
diff --git a/src/intel/compiler/brw_vec4_generator.cpp b/src/intel/compiler/brw_vec4_generator.cpp
index 888cb358fd1..93baaef3ab7 100644
--- a/src/intel/compiler/brw_vec4_generator.cpp
+++ b/src/intel/compiler/brw_vec4_generator.cpp
@@ -291,8 +291,6 @@ generate_tex(struct brw_codegen *p,
inst->header_size != 0,
BRW_SAMPLER_SIMD_MODE_SIMD4X2,
return_format);
-
- brw_mark_surface_used(&prog_data->base, sampler + base_binding_table_index);
} else {
/* Non-constant sampler index. */
@@ -1351,8 +1349,6 @@ generate_get_buffer_size(struct brw_codegen *p,
inst->header_size > 0,
BRW_SAMPLER_SIMD_MODE_SIMD4X2,
BRW_SAMPLER_RETURN_FORMAT_SINT32);
-
- brw_mark_surface_used(&prog_data->base, surf_index.ud);
}
static void
@@ -1378,9 +1374,6 @@ generate_pull_constant_load_gen7(struct brw_codegen *p,
0, /* LD message ignores sampler unit */
GEN5_SAMPLER_MESSAGE_SAMPLE_LD,
BRW_SAMPLER_SIMD_MODE_SIMD4X2, 0));
-
- brw_mark_surface_used(&prog_data->base, surf_index.ud);
-
} else {
struct brw_reg addr = vec1(retype(brw_address_reg(0), BRW_REGISTER_TYPE_UD));
@@ -1866,8 +1859,6 @@ generate_code(struct brw_codegen *p,
case SHADER_OPCODE_SHADER_TIME_ADD:
brw_shader_time_add(p, src[0],
prog_data->base.binding_table.shader_time_start);
- brw_mark_surface_used(&prog_data->base,
- prog_data->base.binding_table.shader_time_start);
break;
case SHADER_OPCODE_UNTYPED_ATOMIC:
diff --git a/src/intel/compiler/brw_vec4_nir.cpp b/src/intel/compiler/brw_vec4_nir.cpp
index f8d4dfb4682..98632b5af08 100644
--- a/src/intel/compiler/brw_vec4_nir.cpp
+++ b/src/intel/compiler/brw_vec4_nir.cpp
@@ -387,16 +387,11 @@ vec4_visitor::get_nir_ssbo_intrinsic_index(nir_intrinsic_instr *instr)
unsigned index = prog_data->base.binding_table.ssbo_start +
nir_src_as_uint(instr->src[src]);
surf_index = brw_imm_ud(index);
- brw_mark_surface_used(&prog_data->base, index);
} else {
surf_index = src_reg(this, glsl_type::uint_type);
emit(ADD(dst_reg(surf_index), get_nir_src(instr->src[src], 1),
brw_imm_ud(prog_data->base.binding_table.ssbo_start)));
surf_index = emit_uniformize(surf_index);
-
- brw_mark_surface_used(&prog_data->base,
- prog_data->base.binding_table.ssbo_start +
- nir->info.num_ssbos - 1);
}
return surf_index;
@@ -492,8 +487,6 @@ vec4_visitor::nir_emit_intrinsic(nir_intrinsic_instr *instr)
emit(MOV(dst_reg(MRF, param_base, glsl_type::int_type, writemask), lod));
emit(inst);
-
- brw_mark_surface_used(&prog_data->base, index);
break;
}
@@ -700,7 +693,6 @@ vec4_visitor::nir_emit_intrinsic(nir_intrinsic_instr *instr)
const unsigned index = prog_data->base.binding_table.ubo_start +
nir_src_as_uint(instr->src[0]);
surf_index = brw_imm_ud(index);
- brw_mark_surface_used(&prog_data->base, index);
} else {
/* The block index is not a constant. Evaluate the index expression
* per-channel and add the base UBO index; we have to select a value
@@ -711,13 +703,6 @@ vec4_visitor::nir_emit_intrinsic(nir_intrinsic_instr *instr)
instr->num_components),
brw_imm_ud(prog_data->base.binding_table.ubo_start)));
surf_index = emit_uniformize(surf_index);
-
- /* Assume this may touch any UBO. It would be nice to provide
- * a tighter bound, but the array information is already lowered away.
- */
- brw_mark_surface_used(&prog_data->base,
- prog_data->base.binding_table.ubo_start +
- nir->info.num_ubos - 1);
}
src_reg offset_reg;
@@ -2031,20 +2016,6 @@ vec4_visitor::nir_emit_texture(nir_tex_instr *instr)
}
case nir_tex_src_texture_offset: {
- /* The highest texture which may be used by this operation is
- * the last element of the array. Mark it here, because the generator
- * doesn't have enough information to determine the bound.
- */
- uint32_t array_size = instr->texture_array_size;
- uint32_t max_used = texture + array_size - 1;
- if (instr->op == nir_texop_tg4) {
- max_used += prog_data->base.binding_table.gather_texture_start;
- } else {
- max_used += prog_data->base.binding_table.texture_start;
- }
-
- brw_mark_surface_used(&prog_data->base, max_used);
-
/* Emit code to evaluate the actual indexing expression */
src_reg src = get_nir_src(instr->src[i].src, 1);
src_reg temp(this, glsl_type::uint_type);
diff --git a/src/intel/compiler/brw_vec4_visitor.cpp b/src/intel/compiler/brw_vec4_visitor.cpp
index aea64c5b54d..a8ff7f4da25 100644
--- a/src/intel/compiler/brw_vec4_visitor.cpp
+++ b/src/intel/compiler/brw_vec4_visitor.cpp
@@ -1748,8 +1748,6 @@ vec4_visitor::emit_pull_constant_load(bblock_t *block, vec4_instruction *inst,
src = byte_offset(src, 16);
}
- brw_mark_surface_used(&prog_data->base, index);
-
if (is_64bit) {
temp = retype(temp, BRW_REGISTER_TYPE_DF);
shuffle_64bit_data(orig_temp, src_reg(temp), false, block, inst);
diff --git a/src/mesa/drivers/dri/i965/brw_nir_uniforms.cpp b/src/mesa/drivers/dri/i965/brw_nir_uniforms.cpp
index 66cdc1a10b6..256fdd8fc79 100644
--- a/src/mesa/drivers/dri/i965/brw_nir_uniforms.cpp
+++ b/src/mesa/drivers/dri/i965/brw_nir_uniforms.cpp
@@ -109,10 +109,6 @@ brw_setup_image_uniform_values(gl_shader_stage stage,
image_idx,
offsetof(brw_image_param, swizzling), 2);
param += BRW_IMAGE_PARAM_SIZE;
-
- brw_mark_surface_used(
- stage_prog_data,
- stage_prog_data->binding_table.image_start + image_idx);
}
}
diff --git a/src/mesa/drivers/dri/i965/brw_program.c b/src/mesa/drivers/dri/i965/brw_program.c
index 730d6dc0d5a..beccd8b3588 100644
--- a/src/mesa/drivers/dri/i965/brw_program.c
+++ b/src/mesa/drivers/dri/i965/brw_program.c
@@ -835,7 +835,10 @@ brw_assign_common_binding_table_offsets(const struct gen_device_info *devinfo,
stage_prog_data->binding_table.plane_start[2] = next_binding_table_offset;
next_binding_table_offset += num_textures;
- /* prog_data->base.binding_table.size will be set by brw_mark_surface_used. */
+ /* Set the binding table size. Some callers may append new entries
+ * and increase this accordingly.
+ */
+ stage_prog_data->binding_table.size_bytes = next_binding_table_offset * 4;
assert(next_binding_table_offset <= BRW_MAX_SURFACES);
return next_binding_table_offset;
diff --git a/src/mesa/drivers/dri/i965/brw_wm.c b/src/mesa/drivers/dri/i965/brw_wm.c
index 536e47638e6..7bbb6166344 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.c
+++ b/src/mesa/drivers/dri/i965/brw_wm.c
@@ -63,6 +63,9 @@ assign_fs_binding_table_offsets(const struct gen_device_info *devinfo,
next_binding_table_offset;
next_binding_table_offset += key->nr_color_regions;
}
+
+ /* Update the binding table size */
+ prog_data->base.binding_table.size_bytes = next_binding_table_offset * 4;
}
static void
--
2.19.1
More information about the mesa-dev
mailing list