- if (stage == MESA_SHADER_FRAGMENT) {
- shader_info->num_input_vgprs = 0;
- if (G_0286CC_PERSP_SAMPLE_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 2;
- if (G_0286CC_PERSP_CENTER_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 2;
- if (G_0286CC_PERSP_CENTROID_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 2;
- if (G_0286CC_PERSP_PULL_MODEL_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 3;
- if (G_0286CC_LINEAR_SAMPLE_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 2;
- if (G_0286CC_LINEAR_CENTER_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 2;
- if (G_0286CC_LINEAR_CENTROID_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 2;
- if (G_0286CC_LINE_STIPPLE_TEX_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 1;
- if (G_0286CC_POS_X_FLOAT_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 1;
- if (G_0286CC_POS_Y_FLOAT_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 1;
- if (G_0286CC_POS_Z_FLOAT_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 1;
- if (G_0286CC_POS_W_FLOAT_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 1;
- if (G_0286CC_FRONT_FACE_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 1;
- if (G_0286CC_ANCILLARY_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 1;
- if (G_0286CC_SAMPLE_COVERAGE_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 1;
- if (G_0286CC_POS_FIXED_PT_ENA(config->spi_ps_input_addr))
- shader_info->num_input_vgprs += 1;
- }
- config->num_vgprs = MAX2(config->num_vgprs, shader_info->num_input_vgprs);
-
- /* +3 for scratch wave offset and VCC */
- config->num_sgprs = MAX2(config->num_sgprs,
- shader_info->num_input_sgprs + 3);
-
- /* Enable 64-bit and 16-bit denormals, because there is no performance
- * cost.
- *
- * If denormals are enabled, all floating-point output modifiers are
- * ignored.
- *
- * Don't enable denormals for 32-bit floats, because:
- * - Floating-point output modifiers would be ignored by the hw.
- * - Some opcodes don't support denormals, such as v_mad_f32. We would
- * have to stop using those.
- * - SI & CI would be very slow.
- */
- config->float_mode |= V_00B028_FP_64_DENORMS;
+ size_t llvm_ir_size = llvm_ir_string ? strlen(llvm_ir_string) : 0;
+ size_t alloc_size = sizeof(struct radv_shader_binary_rtld) + elf_size + llvm_ir_size + 1;
+ struct radv_shader_binary_rtld *rbin = calloc(1, alloc_size);
+ memcpy(rbin->data, elf_buffer, elf_size);
+ if (llvm_ir_string)
+ memcpy(rbin->data + elf_size, llvm_ir_string, llvm_ir_size + 1);
+
+ rbin->base.type = RADV_BINARY_TYPE_RTLD;
+ rbin->base.stage = stage;
+ rbin->base.total_size = alloc_size;
+ rbin->elf_size = elf_size;
+ rbin->llvm_ir_size = llvm_ir_size;
+ *rbinary = &rbin->base;
+
+ free(llvm_ir_string);
+ free(elf_buffer);