[Mesa-dev] [PATCH 08/92] nir: add nir_lower_uniforms_to_ubo pass
Connor Abbott
cwabbott0 at gmail.com
Mon Jun 26 19:07:43 UTC 2017
I think this would probably be a lot simpler if it ran after
nir_lower_io, since you wouldn't need to do all the dereference
munging to compute the offset. That is, instead of translating
load_var intrinsics, it would translate load_uniform intrinsics into
load_ubo intrinsics. Also, it would mean that this would become
entirely a radeonsi-specific thing, instead of having to add a special
hook to the driver-independent state tracker code.
On Mon, Jun 26, 2017 at 7:09 AM, Nicolai Hähnle <nhaehnle at gmail.com> wrote:
> From: Nicolai Hähnle <nicolai.haehnle at amd.com>
>
> This is an alternative lowering of default-block uniforms that doesn't
> use uniform load intrinsics. Instead, it translates the loads to UBO
> loads, allowing a simpler backend.
> ---
> src/compiler/Makefile.sources | 1 +
> src/compiler/nir/nir.h | 1 +
> src/compiler/nir/nir_lower_uniforms_to_ubo.c | 140 +++++++++++++++++++++++++++
> 3 files changed, 142 insertions(+)
> create mode 100644 src/compiler/nir/nir_lower_uniforms_to_ubo.c
>
> diff --git a/src/compiler/Makefile.sources b/src/compiler/Makefile.sources
> index 5f4440b..a0a0d62 100644
> --- a/src/compiler/Makefile.sources
> +++ b/src/compiler/Makefile.sources
> @@ -230,20 +230,21 @@ NIR_FILES = \
> nir/nir_lower_patch_vertices.c \
> nir/nir_lower_phis_to_scalar.c \
> nir/nir_lower_regs_to_ssa.c \
> nir/nir_lower_returns.c \
> nir/nir_lower_samplers.c \
> nir/nir_lower_samplers_as_deref.c \
> nir/nir_lower_system_values.c \
> nir/nir_lower_tex.c \
> nir/nir_lower_to_source_mods.c \
> nir/nir_lower_two_sided_color.c \
> + nir/nir_lower_uniforms_to_ubo.c \
> nir/nir_lower_vars_to_ssa.c \
> nir/nir_lower_var_copies.c \
> nir/nir_lower_vec_to_movs.c \
> nir/nir_lower_wpos_center.c \
> nir/nir_lower_wpos_ytransform.c \
> nir/nir_metadata.c \
> nir/nir_move_vec_src_uses_to_dest.c \
> nir/nir_normalize_cubemap_coords.c \
> nir/nir_opt_conditional_discard.c \
> nir/nir_opt_constant_folding.c \
> diff --git a/src/compiler/nir/nir.h b/src/compiler/nir/nir.h
> index 779bf58..c21ecc5 100644
> --- a/src/compiler/nir/nir.h
> +++ b/src/compiler/nir/nir.h
> @@ -2568,20 +2568,21 @@ void nir_lower_drawpixels(nir_shader *shader,
> typedef struct nir_lower_bitmap_options {
> unsigned sampler;
> bool swizzle_xxxx;
> } nir_lower_bitmap_options;
>
> void nir_lower_bitmap(nir_shader *shader, const nir_lower_bitmap_options *options);
>
> bool nir_lower_atomics(nir_shader *shader,
> const struct gl_shader_program *shader_program);
> bool nir_lower_atomics_to_ssbo(nir_shader *shader, unsigned ssbo_offset);
> +bool nir_lower_uniforms_to_ubo(nir_shader *shader);
> bool nir_lower_to_source_mods(nir_shader *shader);
>
> bool nir_lower_gs_intrinsics(nir_shader *shader);
>
> typedef enum {
> nir_lower_imul64 = (1 << 0),
> nir_lower_isign64 = (1 << 1),
> /** Lower all int64 modulus and division opcodes */
> nir_lower_divmod64 = (1 << 2),
> } nir_lower_int64_options;
> diff --git a/src/compiler/nir/nir_lower_uniforms_to_ubo.c b/src/compiler/nir/nir_lower_uniforms_to_ubo.c
> new file mode 100644
> index 0000000..63f3bc8
> --- /dev/null
> +++ b/src/compiler/nir/nir_lower_uniforms_to_ubo.c
> @@ -0,0 +1,140 @@
> +/*
> + * Copyright 2017 Advanced Micro Devices, Inc.
> + *
> + * Permission is hereby granted, free of charge, to any person obtaining a
> + * copy of this software and associated documentation files (the "Software"),
> + * to deal in the Software without restriction, including without limitation
> + * on the rights to use, copy, modify, merge, publish, distribute, sub
> + * license, and/or sell copies of the Software, and to permit persons to whom
> + * the Software is furnished to do so, subject to the following conditions:
> + *
> + * The above copyright notice and this permission notice (including the next
> + * paragraph) shall be included in all copies or substantial portions of the
> + * Software.
> + *
> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
> + * THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
> + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
> + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
> + * USE OR OTHER DEALINGS IN THE SOFTWARE.
> + */
> +
> +/*
> + * Remap uniform variable accesses to UBO accesses of UBO binding point 0, with
> + * the offset determined by std140 layout, with the base address given by
> + * variable->data.driver_location (in 16-byte units).
> + *
> + * Simultaneously, remap existing UBO accesses by increasing their binding
> + * point by 1.
> + */
> +
> +#include "nir.h"
> +#include "nir_builder.h"
> +
> +static nir_ssa_def *
> +get_deref_offset(nir_deref *tail, nir_ssa_def *offset, nir_builder *b)
> +{
> + while (tail->child != NULL) {
> + const struct glsl_type *parent_type = tail->type;
> + tail = tail->child;
> +
> + if (tail->deref_type == nir_deref_type_array) {
> + nir_deref_array *deref_array = nir_deref_as_array(tail);
> +
> + assert(deref_array->deref_array_type != nir_deref_array_type_wildcard);
> +
> + unsigned size = 16 * glsl_count_attribute_slots(tail->type, false);
> +
> + nir_ssa_def *cur_offset = nir_imm_int(b, deref_array->base_offset);
> + if (deref_array->deref_array_type == nir_deref_array_type_indirect) {
> + cur_offset = nir_iadd(b, cur_offset,
> + nir_ssa_for_src(b, deref_array->indirect, 1));
> + }
> +
> + offset = nir_iadd(b, offset,
> + nir_imul(b, cur_offset, nir_imm_int(b, size)));
> + } else if (tail->deref_type == nir_deref_type_struct) {
> + nir_deref_struct *deref_struct = nir_deref_as_struct(tail);
> +
> + unsigned cur_offset = 0;
> + for (unsigned i = 0; i < deref_struct->index; i++) {
> + const struct glsl_type *ft = glsl_get_struct_field(parent_type, i);
> + cur_offset += 16 * glsl_count_attribute_slots(ft, false);
> + }
> +
> + offset = nir_iadd(b, offset, nir_imm_int(b, cur_offset));
> + }
> + }
> +
> + return offset;
> +}
> +
> +static bool
> +lower_instr(nir_intrinsic_instr *instr, nir_builder *b)
> +{
> + b->cursor = nir_before_instr(&instr->instr);
> +
> + if (instr->intrinsic == nir_intrinsic_load_ubo) {
> + nir_ssa_def *old_idx = nir_ssa_for_src(b, instr->src[0], 1);
> + nir_ssa_def *new_idx = nir_iadd(b, old_idx, nir_imm_int(b, 1));
> + nir_instr_rewrite_src(&instr->instr, &instr->src[0],
> + nir_src_for_ssa(new_idx));
> + return true;
> + }
> +
> + if (instr->intrinsic == nir_intrinsic_load_var) {
> + nir_deref_var *deref = instr->variables[0];
> +
> + if (deref->var->data.mode != nir_var_uniform)
> + return false;
> +
> + nir_ssa_def *ubo_idx = nir_imm_int(b, 0);
> + nir_ssa_def *ubo_offset = nir_imm_int(b, deref->var->data.driver_location * 16);
> +
> + ubo_offset = get_deref_offset(&deref->deref, ubo_offset, b);
> +
> + nir_intrinsic_instr *load =
> + nir_intrinsic_instr_create(b->shader, nir_intrinsic_load_ubo);
> + load->num_components = instr->num_components;
> + load->src[0] = nir_src_for_ssa(ubo_idx);
> + load->src[1] = nir_src_for_ssa(ubo_offset);
> + nir_ssa_dest_init(&load->instr, &load->dest,
> + load->num_components, instr->dest.ssa.bit_size,
> + instr->dest.ssa.name);
> + nir_builder_instr_insert(b, &load->instr);
> + nir_ssa_def_rewrite_uses(&instr->dest.ssa, nir_src_for_ssa(&load->dest.ssa));
> +
> + nir_instr_remove(&instr->instr);
> + return true;
> + }
> +
> + return false;
> +}
> +
> +bool
> +nir_lower_uniforms_to_ubo(nir_shader *shader)
> +{
> + bool progress = false;
> +
> + nir_foreach_function(function, shader) {
> + if (function->impl) {
> + nir_builder builder;
> + nir_builder_init(&builder, function->impl);
> + nir_foreach_block(block, function->impl) {
> + nir_foreach_instr_safe(instr, block) {
> + if (instr->type == nir_instr_type_intrinsic)
> + progress |= lower_instr(nir_instr_as_intrinsic(instr),
> + &builder);
> + }
> + }
> +
> + nir_metadata_preserve(function->impl, nir_metadata_block_index |
> + nir_metadata_dominance);
> + }
> + }
> +
> + return progress;
> +}
> +
> --
> 2.9.3
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
More information about the mesa-dev
mailing list