On Sun, Jan 8, 2017 at 9:26 PM, Kenneth Graunke <kenn...@whitecape.org>
wrote:
> Signed-off-by: Kenneth Graunke <kenn...@whitecape.org>
> ---
> src/intel/vulkan/anv_pipeline.c | 156 ++++++++++++++++++++++++++++++
> +++++++++-
> 1 file changed, 154 insertions(+), 2 deletions(-)
>
> There's one majorly bogus thing here: caching is totally disabled.
> We set key fields based on the NIR, which breaks the dataflow expectations
> for the program cache. A few of those we could look up from the SPIR-V.
> Others...not so much. I'm not sure what to do about it.
>
> Another somewhat bogus thing: we might need to apply the PointMode "it can
> be in either stage" hack to more fields. It's ugly, but unfortunately
> SPIR-V opted not to decide, so we have to look in either place...
>
> diff --git a/src/intel/vulkan/anv_pipeline.c
> b/src/intel/vulkan/anv_pipeline.c
> index b34759a5406..743f82ad69c 100644
> --- a/src/intel/vulkan/anv_pipeline.c
> +++ b/src/intel/vulkan/anv_pipeline.c
> @@ -492,6 +492,151 @@ anv_pipeline_compile_vs(struct anv_pipeline
> *pipeline,
> }
>
> static VkResult
> +anv_pipeline_compile_tcs_tes(struct anv_pipeline *pipeline,
> + struct anv_pipeline_cache *cache,
> + const VkGraphicsPipelineCreateInfo *info,
> + struct anv_shader_module *tcs_module,
> + const char *tcs_entrypoint,
> + const VkSpecializationInfo *tcs_spec_info,
> + struct anv_shader_module *tes_module,
> + const char *tes_entrypoint,
> + const VkSpecializationInfo *tes_spec_info)
> +{
> + const struct gen_device_info *devinfo = &pipeline->device->info;
> + const struct brw_compiler *compiler =
> + pipeline->device->instance->physicalDevice.compiler;
> + struct anv_pipeline_bind_map tcs_map;
> + struct anv_pipeline_bind_map tes_map;
> + struct brw_tcs_prog_key tcs_key = { 0, };
> + struct brw_tes_prog_key tes_key = { 0, };
> + struct anv_shader_bin *tcs_bin = NULL;
> + struct anv_shader_bin *tes_bin = NULL;
> + unsigned char tcs_sha1[20];
> + unsigned char tes_sha1[20];
> +
> + populate_sampler_prog_key(&pipeline->device->info, &tcs_key.tex);
> + populate_sampler_prog_key(&pipeline->device->info, &tes_key.tex);
> + tcs_key.input_vertices = info->pTessellationState->patchControlPoints;
> +
> + if (tcs_module->size > 0 || tes_module->size > 0) {
> + anv_hash_shader(tcs_sha1, &tcs_key, sizeof(tcs_key), tcs_module,
> + tcs_entrypoint, pipeline->layout, tcs_spec_info);
> + anv_hash_shader(tes_sha1, &tes_key, sizeof(tes_key), tes_module,
> + tes_entrypoint, pipeline->layout, tes_spec_info);
> + //tcs_bin = anv_pipeline_cache_search(cache, tcs_sha1, 20);
> + //tes_bin = anv_pipeline_cache_search(cache, tes_sha1, 20);
>
Um... Did you intend to leave these commented out?
> + }
> +
> + if (tcs_bin == NULL || tes_bin == NULL) {
> + struct brw_tcs_prog_data tcs_prog_data = { 0, };
> + struct brw_tes_prog_data tes_prog_data = { 0, };
> + struct anv_pipeline_binding tcs_surface_to_descriptor[256];
> + struct anv_pipeline_binding tcs_sampler_to_descriptor[256];
> + struct anv_pipeline_binding tes_surface_to_descriptor[256];
> + struct anv_pipeline_binding tes_sampler_to_descriptor[256];
> +
> + tcs_map = (struct anv_pipeline_bind_map) {
> + .surface_to_descriptor = tcs_surface_to_descriptor,
> + .sampler_to_descriptor = tcs_sampler_to_descriptor
> + };
> + tes_map = (struct anv_pipeline_bind_map) {
> + .surface_to_descriptor = tes_surface_to_descriptor,
> + .sampler_to_descriptor = tes_sampler_to_descriptor
> + };
> +
> + nir_shader *tcs_nir =
> + anv_pipeline_compile(pipeline, tcs_module, tcs_entrypoint,
> + MESA_SHADER_TESS_CTRL, tcs_spec_info,
> + &tcs_prog_data.base.base, &tcs_map);
> + nir_shader *tes_nir =
> + anv_pipeline_compile(pipeline, tes_module, tes_entrypoint,
> + MESA_SHADER_TESS_EVAL, tes_spec_info,
> + &tes_prog_data.base.base, &tes_map);
> + if (tcs_nir == NULL || tes_nir == NULL)
> + return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
> +
> + nir_lower_tes_patch_vertices(tes_nir,
> tcs_nir->info->tcs.vertices_out);
> +
> + /* The Vulkan 1.0.29 spec, section 21.1 Tessellator says:
> + *
> + * "PointMode. Controls generation of points rather than
> triangles
> + * or lines. This functionality defaults to disabled, and is
> + * enabled if either shader stage includes the execution mode.
> + *
> + * The backend looks at the TES, so OR in the flag from the TCS.
> + */
> + tes_nir->info->tes.point_mode |= tcs_nir->info->tes.point_mode;
>
tes and tcs are in a union in shader_info so if we need point_mode
available in both stages, we need to either add it to both or pull it out
and make it apply to all stages. I think I'd prefer the former but we need
to be more careful here.
> +
> + anv_fill_binding_table(&tcs_prog_data.base.base, 0);
> + anv_fill_binding_table(&tes_prog_data.base.base, 0);
> +
> + void *mem_ctx = ralloc_context(NULL);
> +
> + ralloc_steal(mem_ctx, tcs_nir);
> + ralloc_steal(mem_ctx, tes_nir);
> +
> + tcs_key.tes_primitive_mode = tes_nir->info->tes.primitive_mode;
> + tcs_key.outputs_written = tcs_nir->info->outputs_written;
> + tcs_key.patch_outputs_written = tcs_nir->info->patch_outputs_w
> ritten;
> + tcs_key.quads_workaround =
> + devinfo->gen < 9 &&
> + tes_nir->info->tes.primitive_mode == 7 /* GL_QUADS */ &&
> + tes_nir->info->tes.spacing == TESS_SPACING_EQUAL;
>
Why are all of these in the key if we pull them out of shader_info? Is GL
more dynamic about them?
> +
> + tes_key.inputs_read = tcs_key.outputs_written;
> + tes_key.patch_inputs_read = tcs_key.patch_outputs_written;
> +
> + unsigned code_size;
> + const int shader_time_index = -1;
> + const unsigned *shader_code;
> +
> + shader_code =
> + brw_compile_tcs(compiler, NULL, mem_ctx, &tcs_key,
> &tcs_prog_data,
> + tcs_nir, shader_time_index, &code_size, NULL);
> + if (shader_code == NULL) {
> + ralloc_free(mem_ctx);
> + return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
> + }
> +
> + tcs_bin = anv_pipeline_upload_kernel(pipeline, cache, tcs_sha1, 20,
> + shader_code, code_size,
> + &tcs_prog_data.base.base,
> + sizeof(tcs_prog_data),
> + &tcs_map);
> + if (!tcs_bin) {
> + ralloc_free(mem_ctx);
> + return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
> + }
> +
> + shader_code =
> + brw_compile_tes(compiler, NULL, mem_ctx, &tes_key,
> + &tcs_prog_data.base.vue_map, &tes_prog_data,
> tes_nir,
> + NULL, shader_time_index, &code_size, NULL);
> + if (shader_code == NULL) {
> + ralloc_free(mem_ctx);
> + return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
> + }
> +
> + tes_bin = anv_pipeline_upload_kernel(pipeline, cache, tes_sha1, 20,
> + shader_code, code_size,
> + &tes_prog_data.base.base,
> + sizeof(tes_prog_data),
> + &tes_map);
> + if (!tes_bin) {
> + ralloc_free(mem_ctx);
> + return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
> + }
> +
> + ralloc_free(mem_ctx);
> + }
> +
> + anv_pipeline_add_compiled_stage(pipeline, MESA_SHADER_TESS_CTRL,
> tcs_bin);
> + anv_pipeline_add_compiled_stage(pipeline, MESA_SHADER_TESS_EVAL,
> tes_bin);
> +
> + return VK_SUCCESS;
> +}
> +
> +static VkResult
> anv_pipeline_compile_gs(struct anv_pipeline *pipeline,
> struct anv_pipeline_cache *cache,
> const VkGraphicsPipelineCreateInfo *info,
> @@ -1024,8 +1169,15 @@ anv_pipeline_init(struct anv_pipeline *pipeline,
> goto compile_fail;
> }
>
> - if (modules[MESA_SHADER_TESS_CTRL] || modules[MESA_SHADER_TESS_EVAL])
> - anv_finishme("no tessellation support");
> + if (modules[MESA_SHADER_TESS_EVAL]) {
> + anv_pipeline_compile_tcs_tes(pipeline, cache, pCreateInfo,
> + modules[MESA_SHADER_TESS_CTRL],
> + pStages[MESA_SHADER_TESS_CTRL]->pName,
> + pStages[MESA_SHADER_TESS_
> CTRL]->pSpecializationInfo,
> + modules[MESA_SHADER_TESS_EVAL],
> + pStages[MESA_SHADER_TESS_EVAL]->pName,
> + pStages[MESA_SHADER_TESS_
> EVAL]->pSpecializationInfo);
> + }
>
> if (modules[MESA_SHADER_GEOMETRY]) {
> result = anv_pipeline_compile_gs(pipeline, cache, pCreateInfo,
> --
> 2.11.0
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
