summaryrefslogtreecommitdiffstats
path: root/src/amd/common
diff options
context:
space:
mode:
Diffstat (limited to 'src/amd/common')
-rw-r--r--src/amd/common/ac_llvm_build.c42
-rw-r--r--src/amd/common/ac_llvm_build.h14
-rw-r--r--src/amd/common/ac_nir_to_llvm.c22
3 files changed, 42 insertions, 36 deletions
diff --git a/src/amd/common/ac_llvm_build.c b/src/amd/common/ac_llvm_build.c
index 1d97b09cd4f..949f181aace 100644
--- a/src/amd/common/ac_llvm_build.c
+++ b/src/amd/common/ac_llvm_build.c
@@ -717,32 +717,40 @@ ac_build_indexed_store(struct ac_llvm_context *ctx,
* \param base_ptr Where the array starts.
* \param index The element index into the array.
* \param uniform Whether the base_ptr and index can be assumed to be
- * dynamically uniform
+ * dynamically uniform (i.e. load to an SGPR)
+ * \param invariant Whether the load is invariant (no other opcodes affect it)
*/
-LLVMValueRef
-ac_build_indexed_load(struct ac_llvm_context *ctx,
- LLVMValueRef base_ptr, LLVMValueRef index,
- bool uniform)
+static LLVMValueRef
+ac_build_load_custom(struct ac_llvm_context *ctx, LLVMValueRef base_ptr,
+ LLVMValueRef index, bool uniform, bool invariant)
{
- LLVMValueRef pointer;
+ LLVMValueRef pointer, result;
pointer = ac_build_gep0(ctx, base_ptr, index);
if (uniform)
LLVMSetMetadata(pointer, ctx->uniform_md_kind, ctx->empty_md);
- return LLVMBuildLoad(ctx->builder, pointer, "");
+ result = LLVMBuildLoad(ctx->builder, pointer, "");
+ if (invariant)
+ LLVMSetMetadata(result, ctx->invariant_load_md_kind, ctx->empty_md);
+ return result;
}
-/**
- * Do a load from &base_ptr[index], but also add a flag that it's loading
- * a constant from a dynamically uniform index.
- */
-LLVMValueRef
-ac_build_indexed_load_const(struct ac_llvm_context *ctx,
- LLVMValueRef base_ptr, LLVMValueRef index)
+LLVMValueRef ac_build_load(struct ac_llvm_context *ctx, LLVMValueRef base_ptr,
+ LLVMValueRef index)
{
- LLVMValueRef result = ac_build_indexed_load(ctx, base_ptr, index, true);
- LLVMSetMetadata(result, ctx->invariant_load_md_kind, ctx->empty_md);
- return result;
+ return ac_build_load_custom(ctx, base_ptr, index, false, false);
+}
+
+LLVMValueRef ac_build_load_invariant(struct ac_llvm_context *ctx,
+ LLVMValueRef base_ptr, LLVMValueRef index)
+{
+ return ac_build_load_custom(ctx, base_ptr, index, false, true);
+}
+
+LLVMValueRef ac_build_load_to_sgpr(struct ac_llvm_context *ctx,
+ LLVMValueRef base_ptr, LLVMValueRef index)
+{
+ return ac_build_load_custom(ctx, base_ptr, index, true, true);
}
/* TBUFFER_STORE_FORMAT_{X,XY,XYZ,XYZW} <- the suffix is selected by num_channels=1..4.
diff --git a/src/amd/common/ac_llvm_build.h b/src/amd/common/ac_llvm_build.h
index ac8ea9c4152..f0b5875b423 100644
--- a/src/amd/common/ac_llvm_build.h
+++ b/src/amd/common/ac_llvm_build.h
@@ -150,14 +150,12 @@ ac_build_indexed_store(struct ac_llvm_context *ctx,
LLVMValueRef base_ptr, LLVMValueRef index,
LLVMValueRef value);
-LLVMValueRef
-ac_build_indexed_load(struct ac_llvm_context *ctx,
- LLVMValueRef base_ptr, LLVMValueRef index,
- bool uniform);
-
-LLVMValueRef
-ac_build_indexed_load_const(struct ac_llvm_context *ctx,
- LLVMValueRef base_ptr, LLVMValueRef index);
+LLVMValueRef ac_build_load(struct ac_llvm_context *ctx, LLVMValueRef base_ptr,
+ LLVMValueRef index);
+LLVMValueRef ac_build_load_invariant(struct ac_llvm_context *ctx,
+ LLVMValueRef base_ptr, LLVMValueRef index);
+LLVMValueRef ac_build_load_to_sgpr(struct ac_llvm_context *ctx,
+ LLVMValueRef base_ptr, LLVMValueRef index);
void
ac_build_buffer_store_dword(struct ac_llvm_context *ctx,
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 11ba487f375..3ba3ebf051e 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common/ac_nir_to_llvm.c
@@ -798,7 +798,7 @@ static void create_function(struct nir_to_llvm_context *ctx)
for (unsigned i = 0; i < num_sets; ++i) {
if (ctx->options->layout->set[i].layout->shader_stages & (1 << ctx->stage)) {
set_userdata_location_indirect(&ctx->shader_info->user_sgprs_locs.descriptor_sets[i], desc_sgpr_idx, 2, i * 8);
- ctx->descriptor_sets[i] = ac_build_indexed_load_const(&ctx->ac, desc_sets, LLVMConstInt(ctx->i32, i, false));
+ ctx->descriptor_sets[i] = ac_build_load_to_sgpr(&ctx->ac, desc_sets, LLVMConstInt(ctx->i32, i, false));
} else
ctx->descriptor_sets[i] = NULL;
@@ -2524,7 +2524,7 @@ lds_load(struct nir_to_llvm_context *ctx,
LLVMValueRef dw_addr)
{
LLVMValueRef value;
- value = ac_build_indexed_load(&ctx->ac, ctx->lds, dw_addr, false);
+ value = ac_build_load(&ctx->ac, ctx->lds, dw_addr);
return value;
}
@@ -3739,7 +3739,7 @@ static LLVMValueRef load_sample_position(struct nir_to_llvm_context *ctx,
const_array(ctx->v2f32, 64), "");
sample_id = LLVMBuildAdd(ctx->builder, sample_id, ctx->sample_pos_offset, "");
- result = ac_build_indexed_load(&ctx->ac, ptr, sample_id, false);
+ result = ac_build_load_invariant(&ctx->ac, ptr, sample_id);
return result;
}
@@ -4230,7 +4230,7 @@ static LLVMValueRef radv_get_sampler_desc(struct ac_shader_abi *abi,
list = ac_build_gep0(&ctx->ac, list, LLVMConstInt(ctx->i32, offset, 0));
list = LLVMBuildPointerCast(builder, list, const_array(type, 0), "");
- return ac_build_indexed_load_const(&ctx->ac, list, index);
+ return ac_build_load_to_sgpr(&ctx->ac, list, index);
}
static LLVMValueRef get_sampler_desc(struct ac_nir_context *ctx,
@@ -4927,7 +4927,7 @@ handle_vs_input_decl(struct nir_to_llvm_context *ctx,
for (unsigned i = 0; i < attrib_count; ++i, ++idx) {
t_offset = LLVMConstInt(ctx->i32, index + i, false);
- t_list = ac_build_indexed_load_const(&ctx->ac, t_list_ptr, t_offset);
+ t_list = ac_build_load_to_sgpr(&ctx->ac, t_list_ptr, t_offset);
input = ac_build_buffer_load_format(&ctx->ac, t_list,
buffer_index,
@@ -6217,16 +6217,16 @@ ac_setup_rings(struct nir_to_llvm_context *ctx)
{
if ((ctx->stage == MESA_SHADER_VERTEX && ctx->options->key.vs.as_es) ||
(ctx->stage == MESA_SHADER_TESS_EVAL && ctx->options->key.tes.as_es)) {
- ctx->esgs_ring = ac_build_indexed_load_const(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_ESGS_VS, false));
+ ctx->esgs_ring = ac_build_load_to_sgpr(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_ESGS_VS, false));
}
if (ctx->is_gs_copy_shader) {
- ctx->gsvs_ring = ac_build_indexed_load_const(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_GSVS_VS, false));
+ ctx->gsvs_ring = ac_build_load_to_sgpr(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_GSVS_VS, false));
}
if (ctx->stage == MESA_SHADER_GEOMETRY) {
LLVMValueRef tmp;
- ctx->esgs_ring = ac_build_indexed_load_const(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_ESGS_GS, false));
- ctx->gsvs_ring = ac_build_indexed_load_const(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_GSVS_GS, false));
+ ctx->esgs_ring = ac_build_load_to_sgpr(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_ESGS_GS, false));
+ ctx->gsvs_ring = ac_build_load_to_sgpr(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_GSVS_GS, false));
ctx->gsvs_ring = LLVMBuildBitCast(ctx->builder, ctx->gsvs_ring, ctx->v4i32, "");
@@ -6238,8 +6238,8 @@ ac_setup_rings(struct nir_to_llvm_context *ctx)
if (ctx->stage == MESA_SHADER_TESS_CTRL ||
ctx->stage == MESA_SHADER_TESS_EVAL) {
- ctx->hs_ring_tess_offchip = ac_build_indexed_load_const(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_HS_TESS_OFFCHIP, false));
- ctx->hs_ring_tess_factor = ac_build_indexed_load_const(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_HS_TESS_FACTOR, false));
+ ctx->hs_ring_tess_offchip = ac_build_load_to_sgpr(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_HS_TESS_OFFCHIP, false));
+ ctx->hs_ring_tess_factor = ac_build_load_to_sgpr(&ctx->ac, ctx->ring_offsets, LLVMConstInt(ctx->i32, RING_HS_TESS_FACTOR, false));
}
}
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-24 13:30:09 +0000