ac: add si_nir_load_input_gs() to the abi

V2: make use of driver_location and don't expose NIR to the ABI.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index db1936b..96ba289 100644 (file)
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common/ac_nir_to_llvm.c
@@ -2894,27 +2894,29 @@ load_tes_input(struct nir_to_llvm_context *ctx,
 }
 
 static LLVMValueRef
-load_gs_input(struct nir_to_llvm_context *ctx,
-             nir_intrinsic_instr *instr)
+load_gs_input(struct ac_shader_abi *abi,
+             unsigned location,
+             unsigned driver_location,
+             unsigned component,
+             unsigned num_components,
+             unsigned vertex_index,
+             unsigned const_index,
+             LLVMTypeRef type)
 {
-       LLVMValueRef indir_index, vtx_offset;
-       unsigned const_index;
+       struct nir_to_llvm_context *ctx = nir_to_llvm_context_from_abi(abi);
+       LLVMValueRef vtx_offset;
        LLVMValueRef args[9];
        unsigned param, vtx_offset_param;
        LLVMValueRef value[4], result;
-       unsigned vertex_index;
-       get_deref_offset(ctx->nir, instr->variables[0],
-                        false, &vertex_index, NULL,
-                        &const_index, &indir_index);
+
        vtx_offset_param = vertex_index;
        assert(vtx_offset_param < 6);
        vtx_offset = LLVMBuildMul(ctx->builder, ctx->gs_vtx_offset[vtx_offset_param],
                                  LLVMConstInt(ctx->ac.i32, 4, false), "");
 
-       param = shader_io_get_unique_index(instr->variables[0]->var->data.location);
+       param = shader_io_get_unique_index(location);
 
-       unsigned comp = instr->variables[0]->var->data.location_frac;
-       for (unsigned i = comp; i < instr->num_components + comp; i++) {
+       for (unsigned i = component; i < num_components + component; i++) {
                if (ctx->ac.chip_class >= GFX9) {
                        LLVMValueRef dw_addr = ctx->gs_vtx_offset[vtx_offset_param];
                        dw_addr = LLVMBuildAdd(ctx->ac.builder, dw_addr,
@@ -2937,7 +2939,7 @@ load_gs_input(struct nir_to_llvm_context *ctx,
                                                      AC_FUNC_ATTR_LEGACY);
                }
        }
-       result = ac_build_varying_gather_values(&ctx->ac, value, instr->num_components, comp);
+       result = ac_build_varying_gather_values(&ctx->ac, value, num_components, component);
 
        return result;
 }
@@ -3006,7 +3008,16 @@ static LLVMValueRef visit_load_var(struct ac_nir_context *ctx,
                if (ctx->stage == MESA_SHADER_TESS_EVAL)
                        return load_tes_input(ctx->nctx, instr);
                if (ctx->stage == MESA_SHADER_GEOMETRY) {
-                       return load_gs_input(ctx->nctx, instr);
+                               LLVMValueRef indir_index;
+                               unsigned const_index, vertex_index;
+                               get_deref_offset(ctx, instr->variables[0],
+                                                false, &vertex_index, NULL,
+                                                &const_index, &indir_index);
+                       return ctx->abi->load_inputs(ctx->abi, instr->variables[0]->var->data.location,
+                                                    instr->variables[0]->var->data.driver_location,
+                                                    instr->variables[0]->var->data.location_frac, ve,
+                                                    vertex_index, const_index,
+                                                    nir2llvmtype(ctx, instr->variables[0]->var->type));
                }
 
                for (unsigned chan = comp; chan < ve + comp; chan++) {
@@ -6560,8 +6571,8 @@ LLVMModuleRef ac_translate_nir_to_llvm(LLVMTargetMachineRef tm,
 
                if (shaders[i]->info.stage == MESA_SHADER_GEOMETRY) {
                        ctx.gs_next_vertex = ac_build_alloca(&ctx.ac, ctx.ac.i32, "gs_next_vertex");
-
                        ctx.gs_max_out_vertices = shaders[i]->info.gs.vertices_out;
+                       ctx.abi.load_inputs = load_gs_input;
                } 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;

diff --git a/src/amd/common/ac_shader_abi.h b/src/amd/common/ac_shader_abi.h
index 56209bd..68fc431 100644 (file)
--- a/src/amd/common/ac_shader_abi.h
+++ b/src/amd/common/ac_shader_abi.h
@@ -64,6 +64,15 @@ struct ac_shader_abi {
                            unsigned stream,
                            LLVMValueRef *addrs);
 
+       LLVMValueRef (*load_inputs)(struct ac_shader_abi *abi,
+                                   unsigned location,
+                                   unsigned driver_location,
+                                   unsigned component,
+                                   unsigned num_components,
+                                   unsigned vertex_index,
+                                   unsigned const_index,
+                                   LLVMTypeRef type);
+
        LLVMValueRef (*load_ubo)(struct ac_shader_abi *abi, LLVMValueRef index);
 
        /**

diff --git a/src/gallium/drivers/radeonsi/si_shader.c b/src/gallium/drivers/radeonsi/si_shader.c
index 950207b..a94c2af 100644 (file)
--- a/src/gallium/drivers/radeonsi/si_shader.c
+++ b/src/gallium/drivers/radeonsi/si_shader.c
@@ -5800,6 +5800,7 @@ static bool si_compile_tgsi_main(struct si_shader_context *ctx,
                break;
        case PIPE_SHADER_GEOMETRY:
                bld_base->emit_fetch_funcs[TGSI_FILE_INPUT] = fetch_input_gs;
+               ctx->abi.load_inputs = si_nir_load_input_gs;
                ctx->abi.emit_vertex = si_llvm_emit_vertex;
                ctx->abi.emit_outputs = si_llvm_emit_gs_epilogue;
                bld_base->emit_epilogue = si_tgsi_emit_gs_epilogue;

diff --git a/src/gallium/drivers/radeonsi/si_shader_internal.h b/src/gallium/drivers/radeonsi/si_shader_internal.h
index a548cf1..f50a022 100644 (file)
--- a/src/gallium/drivers/radeonsi/si_shader_internal.h
+++ b/src/gallium/drivers/radeonsi/si_shader_internal.h
@@ -332,4 +332,13 @@ void si_llvm_load_input_fs(
 
 bool si_nir_build_llvm(struct si_shader_context *ctx, struct nir_shader *nir);
 
+LLVMValueRef si_nir_load_input_gs(struct ac_shader_abi *abi,
+                                 unsigned location,
+                                 unsigned driver_location,
+                                 unsigned component,
+                                 unsigned num_components,
+                                 unsigned vertex_index,
+                                 unsigned const_index,
+                                 LLVMTypeRef type);
+
 #endif

diff --git a/src/gallium/drivers/radeonsi/si_shader_nir.c b/src/gallium/drivers/radeonsi/si_shader_nir.c
index 979361a..1b502b3 100644 (file)
--- a/src/gallium/drivers/radeonsi/si_shader_nir.c
+++ b/src/gallium/drivers/radeonsi/si_shader_nir.c
@@ -488,6 +488,26 @@ static void declare_nir_input_fs(struct si_shader_context *ctx,
        si_llvm_load_input_fs(ctx, input_index, out);
 }
 
+LLVMValueRef si_nir_load_input_gs(struct ac_shader_abi *abi,
+                                 unsigned location,
+                                 unsigned driver_location,
+                                 unsigned component,
+                                 unsigned num_components,
+                                 unsigned vertex_index,
+                                 unsigned const_index,
+                                 LLVMTypeRef type)
+{
+       struct si_shader_context *ctx = si_shader_context_from_abi(abi);
+
+       LLVMValueRef value[4];
+       for (unsigned i = component; i < num_components + component; i++) {
+               value[i] = si_llvm_load_input_gs(&ctx->abi, driver_location  / 4,
+                                                vertex_index, type, i);
+       }
+
+       return ac_build_varying_gather_values(&ctx->ac, value, num_components, component);
+}
+
 static LLVMValueRef
 si_nir_load_sampler_desc(struct ac_shader_abi *abi,
                         unsigned descriptor_set, unsigned base_index,