On 12.01.2018 03:55, Timothy Arceri wrote:
This shares mode code and calls the new shared load_tess_varyings()
s/mode/more/?
abi so that the radeonsi nir path now supports tcs output loads.
---
src/amd/common/ac_nir_to_llvm.c | 126 +++++++++++++++++-----------------------
1 file changed, 53 insertions(+), 73 deletions(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 984a8f531d..8d759b30fb 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common/ac_nir_to_llvm.c
@@ -2807,73 +2807,45 @@ get_dw_address(struct nir_to_llvm_context *ctx,
}
static LLVMValueRef
-load_tcs_input(struct ac_shader_abi *abi,
- LLVMValueRef vertex_index,
- LLVMValueRef indir_index,
- unsigned const_index,
- unsigned location,
- unsigned driver_location,
- unsigned component,
- unsigned num_components,
- bool is_patch,
- bool is_compact,
- bool load_input)
+load_tcs_varyings(struct ac_shader_abi *abi,
+ LLVMValueRef vertex_index,
+ LLVMValueRef indir_index,
+ unsigned const_index,
+ unsigned location,
+ unsigned driver_location,
+ unsigned component,
+ unsigned num_components,
+ bool is_patch,
+ bool is_compact,
+ bool load_input)
{
struct nir_to_llvm_context *ctx = nir_to_llvm_context_from_abi(abi);
LLVMValueRef dw_addr, stride;
LLVMValueRef value[4], result;
unsigned param = shader_io_get_unique_index(location);
- stride = unpack_param(&ctx->ac, ctx->tcs_in_layout, 13, 8);
- dw_addr = get_tcs_in_current_patch_offset(ctx);
- dw_addr = get_dw_address(ctx, dw_addr, param, const_index, is_compact,
vertex_index, stride,
- indir_index);
-
- for (unsigned i = 0; i < num_components + component; i++) {
- value[i] = ac_lds_load(&ctx->ac, dw_addr);
- dw_addr = LLVMBuildAdd(ctx->builder, dw_addr,
- ctx->ac.i32_1, "");
- }
- result = ac_build_varying_gather_values(&ctx->ac, value,
num_components, component);
- return result;
-}
-
-static LLVMValueRef
-load_tcs_output(struct nir_to_llvm_context *ctx,
- nir_intrinsic_instr *instr)
-{
- LLVMValueRef dw_addr;
- LLVMValueRef stride = NULL;
- LLVMValueRef value[4], result;
- LLVMValueRef vertex_index = NULL;
- LLVMValueRef indir_index = NULL;
- unsigned const_index = 0;
- unsigned param;
- const bool per_vertex = nir_is_per_vertex_io(instr->variables[0]->var,
ctx->stage);
- const bool is_compact = instr->variables[0]->var->data.compact;
- param =
shader_io_get_unique_index(instr->variables[0]->var->data.location);
- get_deref_offset(ctx->nir, instr->variables[0],
- false, NULL, per_vertex ? &vertex_index : NULL,
- &const_index, &indir_index);
-
- if (!instr->variables[0]->var->data.patch) {
- stride = unpack_param(&ctx->ac, ctx->tcs_out_layout, 13, 8);
- dw_addr = get_tcs_out_current_patch_offset(ctx);
+ if (load_input) {
+ stride = unpack_param(&ctx->ac, ctx->tcs_in_layout, 13, 8);
+ dw_addr = get_tcs_in_current_patch_offset(ctx);
} else {
- dw_addr = get_tcs_out_current_patch_data_offset(ctx);
+ if (!is_patch) {
+ stride = unpack_param(&ctx->ac, ctx->tcs_out_layout,
13, 8);
+ dw_addr = get_tcs_out_current_patch_offset(ctx);
+ } else {
+ dw_addr = get_tcs_out_current_patch_data_offset(ctx);
+ stride = NULL;
+ }
}
dw_addr = get_dw_address(ctx, dw_addr, param, const_index, is_compact, vertex_index, stride,
indir_index);
- unsigned comp = instr->variables[0]->var->data.location_frac;
- for (unsigned i = comp; i < instr->num_components + comp; i++) {
+ for (unsigned i = 0; i < num_components + component; i++) {
value[i] = ac_lds_load(&ctx->ac, dw_addr);
dw_addr = LLVMBuildAdd(ctx->builder, dw_addr,
ctx->ac.i32_1, "");
}
So, I realize the code was like this in the load_tcs_input path before,
and the load_tcs_output path was apparently incorrect, but it would
still be nicer and less confusing if we didn't call ac_lds_load for i <
component, as those values will just be unused anyway.
Cheers,
Nicolai
- result = ac_build_varying_gather_values(&ctx->ac, value,
instr->num_components, comp);
- result = LLVMBuildBitCast(ctx->builder, result, get_def_type(ctx->nir,
&instr->dest.ssa), "");
+ result = ac_build_varying_gather_values(&ctx->ac, value,
num_components, component);
return result;
}
@@ -3081,6 +3053,31 @@ build_gep_for_deref(struct ac_nir_context *ctx,
return val;
}
+static LLVMValueRef load_tess_varyings(struct ac_nir_context *ctx,
+ nir_intrinsic_instr *instr,
+ bool load_inputs)
+{
+ LLVMValueRef result;
+ LLVMValueRef vertex_index = NULL;
+ LLVMValueRef indir_index = NULL;
+ unsigned const_index = 0;
+ unsigned location = instr->variables[0]->var->data.location;
+ unsigned driver_location =
instr->variables[0]->var->data.driver_location;
+ const bool is_patch = instr->variables[0]->var->data.patch;
+ const bool is_compact = instr->variables[0]->var->data.compact;
+
+ get_deref_offset(ctx, instr->variables[0],
+ false, NULL, is_patch ? NULL : &vertex_index,
+ &const_index, &indir_index);
+
+ result = ctx->abi->load_tess_varyings(ctx->abi, vertex_index,
indir_index,
+ const_index, location,
driver_location,
+
instr->variables[0]->var->data.location_frac,
+ instr->num_components,
+ is_patch, is_compact,
load_inputs);
+ return LLVMBuildBitCast(ctx->ac.builder, result, get_def_type(ctx,
&instr->dest.ssa), "");
+}
+
static LLVMValueRef visit_load_var(struct ac_nir_context *ctx,
nir_intrinsic_instr *instr)
{
@@ -3104,25 +3101,7 @@ static LLVMValueRef visit_load_var(struct ac_nir_context
*ctx,
case nir_var_shader_in:
if (ctx->stage == MESA_SHADER_TESS_CTRL ||
ctx->stage == MESA_SHADER_TESS_EVAL) {
- LLVMValueRef result;
- LLVMValueRef vertex_index = NULL;
- LLVMValueRef indir_index = NULL;
- unsigned const_index = 0;
- unsigned location =
instr->variables[0]->var->data.location;
- unsigned driver_location =
instr->variables[0]->var->data.driver_location;
- const bool is_patch =
instr->variables[0]->var->data.patch;
- const bool is_compact =
instr->variables[0]->var->data.compact;
-
- get_deref_offset(ctx, instr->variables[0],
- false, NULL, is_patch ? NULL :
&vertex_index,
- &const_index, &indir_index);
-
- result = ctx->abi->load_tess_varyings(ctx->abi,
vertex_index, indir_index,
- const_index,
location, driver_location,
-
instr->variables[0]->var->data.location_frac,
-
instr->num_components,
- is_patch,
is_compact, true);
- return LLVMBuildBitCast(ctx->ac.builder, result, get_def_type(ctx,
&instr->dest.ssa), "");
+ return load_tess_varyings(ctx, instr, true);
}
if (ctx->stage == MESA_SHADER_GEOMETRY) {
@@ -3182,8 +3161,9 @@ static LLVMValueRef visit_load_var(struct ac_nir_context
*ctx,
"");
}
case nir_var_shader_out:
- if (ctx->stage == MESA_SHADER_TESS_CTRL)
- return load_tcs_output(ctx->nctx, instr);
+ if (ctx->stage == MESA_SHADER_TESS_CTRL) {
+ return load_tess_varyings(ctx, instr, false);
+ }
for (unsigned chan = comp; chan < ve + comp; chan++) {
if (indir_index) {
@@ -6706,7 +6686,7 @@ LLVMModuleRef
ac_translate_nir_to_llvm(LLVMTargetMachineRef tm,
} else if (shaders[i]->info.stage == MESA_SHADER_TESS_CTRL) {
ctx.tcs_outputs_read = shaders[i]->info.outputs_read;
ctx.tcs_patch_outputs_read =
shaders[i]->info.patch_outputs_read;
- ctx.abi.load_tess_varyings = load_tcs_input;
+ ctx.abi.load_tess_varyings = load_tcs_varyings;
ctx.abi.load_patch_vertices_in = load_patch_vertices_in;
ctx.abi.store_tcs_outputs = store_tcs_output;
} else if (shaders[i]->info.stage == MESA_SHADER_TESS_EVAL) {
--
Lerne, wie die Welt wirklich ist,
Aber vergiss niemals, wie sie sein sollte.
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
