radv: Fix threading issue with submission refcounts.

[mesa.git] / src / amd / vulkan / radv_meta_decompress.c

diff --git a/src/amd/vulkan/radv_meta_decompress.c b/src/amd/vulkan/radv_meta_decompress.c
index 43412f69a639dd04c809831722473b53546afa57..5ab7e4fe243d375e809a206e30debc67cc090981 100644 (file)
--- a/src/amd/vulkan/radv_meta_decompress.c
+++ b/src/amd/vulkan/radv_meta_decompress.c
@@ -125,7 +125,6 @@ create_pipeline_layout(struct radv_device *device, VkPipelineLayout *layout)
 
 static VkResult
 create_pipeline(struct radv_device *device,
-                VkShaderModule vs_module_h,
                uint32_t samples,
                VkRenderPass pass,
                VkPipelineLayout layout,
@@ -135,7 +134,6 @@ create_pipeline(struct radv_device *device,
 {
        VkResult result;
        VkDevice device_h = radv_device_to_handle(device);
-       struct radv_shader_module vs_module = {0};
 
        mtx_lock(&device->meta_state.mtx);
        if (*pipeline) {
@@ -143,9 +141,14 @@ create_pipeline(struct radv_device *device,
                return VK_SUCCESS;
        }
 
-       if (!vs_module_h) {
-               vs_module.nir = radv_meta_build_nir_vs_generate_vertices();
-               vs_module_h = radv_shader_module_to_handle(&vs_module);
+       struct radv_shader_module vs_module = {
+               .nir = radv_meta_build_nir_vs_generate_vertices()
+       };
+
+       if (!vs_module.nir) {
+               /* XXX: Need more accurate error */
+               result = VK_ERROR_OUT_OF_HOST_MEMORY;
+               goto cleanup;
        }
 
        struct radv_shader_module fs_module = {
@@ -170,7 +173,7 @@ create_pipeline(struct radv_device *device,
                       {
                                .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO,
                                .stage = VK_SHADER_STAGE_VERTEX_BIT,
-                               .module = vs_module_h,
+                               .module = radv_shader_module_to_handle(&vs_module),
                                .pName = "main",
                        },
                        {
@@ -241,11 +244,11 @@ create_pipeline(struct radv_device *device,
 
        struct radv_graphics_pipeline_create_info extra = {
                .use_rectlist = true,
-               .db_flush_depth_inplace = decompress == DECOMPRESS_DEPTH_STENCIL ||
+               .depth_compress_disable = decompress == DECOMPRESS_DEPTH_STENCIL ||
                                          decompress == DECOMPRESS_DEPTH,
-               .db_flush_stencil_inplace = decompress == DECOMPRESS_DEPTH_STENCIL ||
+               .stencil_compress_disable = decompress == DECOMPRESS_DEPTH_STENCIL ||
                                            decompress == DECOMPRESS_STENCIL,
-               .db_resummarize = op == DEPTH_RESUMMARIZE,
+               .resummarize_enable = op == DEPTH_RESUMMARIZE,
        };
 
        result = radv_graphics_pipeline_create(device_h,
@@ -256,8 +259,7 @@ create_pipeline(struct radv_device *device,
 
 cleanup:
        ralloc_free(fs_module.nir);
-       if (vs_module.nir)
-               ralloc_free(vs_module.nir);
+       ralloc_free(vs_module.nir);
        mtx_unlock(&device->meta_state.mtx);
        return result;
 }
@@ -292,15 +294,6 @@ radv_device_init_meta_depth_decomp_state(struct radv_device *device, bool on_dem
        struct radv_meta_state *state = &device->meta_state;
        VkResult res = VK_SUCCESS;
 
-       struct radv_shader_module vs_module = { .nir = radv_meta_build_nir_vs_generate_vertices() };
-       if (!vs_module.nir) {
-               /* XXX: Need more accurate error */
-               res = VK_ERROR_OUT_OF_HOST_MEMORY;
-               goto fail;
-       }
-
-       VkShaderModule vs_module_h = radv_shader_module_to_handle(&vs_module);
-
        for (uint32_t i = 0; i < ARRAY_SIZE(state->depth_decomp); ++i) {
                uint32_t samples = 1 << i;
 
@@ -317,7 +310,7 @@ radv_device_init_meta_depth_decomp_state(struct radv_device *device, bool on_dem
                        continue;
 
                for (uint32_t j = 0; j < NUM_DEPTH_DECOMPRESS_PIPELINES; j++) {
-                       res = create_pipeline(device, vs_module_h, samples,
+                       res = create_pipeline(device, samples,
                                              state->depth_decomp[i].pass,
                                              state->depth_decomp[i].p_layout,
                                              DEPTH_DECOMPRESS,
@@ -327,7 +320,7 @@ radv_device_init_meta_depth_decomp_state(struct radv_device *device, bool on_dem
                                goto fail;
                }
 
-               res = create_pipeline(device, vs_module_h, samples,
+               res = create_pipeline(device, samples,
                                      state->depth_decomp[i].pass,
                                      state->depth_decomp[i].p_layout,
                                      DEPTH_RESUMMARIZE,
@@ -337,14 +330,10 @@ radv_device_init_meta_depth_decomp_state(struct radv_device *device, bool on_dem
                        goto fail;
        }
 
-       goto cleanup;
+       return VK_SUCCESS;
 
 fail:
        radv_device_finish_meta_depth_decomp_state(device);
-
-cleanup:
-       ralloc_free(vs_module.nir);
-
        return res;
 }
 
@@ -372,7 +361,7 @@ radv_get_depth_pipeline(struct radv_cmd_buffer *cmd_buffer,
                VkResult ret;
 
                for (uint32_t i = 0; i < NUM_DEPTH_DECOMPRESS_PIPELINES; i++) {
-                       ret = create_pipeline(cmd_buffer->device, VK_NULL_HANDLE, samples,
+                       ret = create_pipeline(cmd_buffer->device, samples,
                                              state->depth_decomp[samples_log2].pass,
                                              state->depth_decomp[samples_log2].p_layout,
                                              DEPTH_DECOMPRESS,
@@ -384,7 +373,7 @@ radv_get_depth_pipeline(struct radv_cmd_buffer *cmd_buffer,
                        }
                }
 
-               ret = create_pipeline(cmd_buffer->device, VK_NULL_HANDLE, samples,
+               ret = create_pipeline(cmd_buffer->device, samples,
                                      state->depth_decomp[samples_log2].pass,
                                      state->depth_decomp[samples_log2].p_layout,
                                      DEPTH_RESUMMARIZE,
@@ -482,11 +471,11 @@ radv_process_depth_image_layer(struct radv_cmd_buffer *cmd_buffer,
                                &cmd_buffer->pool->alloc);
 }
 
-static void radv_process_depth_image_inplace(struct radv_cmd_buffer *cmd_buffer,
-                                            struct radv_image *image,
-                                            const VkImageSubresourceRange *subresourceRange,
-                                            struct radv_sample_locations_state *sample_locs,
-                                            enum radv_depth_op op)
+static void radv_process_depth_stencil(struct radv_cmd_buffer *cmd_buffer,
+                                      struct radv_image *image,
+                                      const VkImageSubresourceRange *subresourceRange,
+                                      struct radv_sample_locations_state *sample_locs,
+                                      enum radv_depth_op op)
 {
        struct radv_meta_saved_state saved_state;
        VkCommandBuffer cmd_buffer_h = radv_cmd_buffer_to_handle(cmd_buffer);
@@ -554,22 +543,32 @@ static void radv_process_depth_image_inplace(struct radv_cmd_buffer *cmd_buffer,
        radv_meta_restore(&saved_state, cmd_buffer);
 }
 
-void radv_decompress_depth_image_inplace(struct radv_cmd_buffer *cmd_buffer,
-                                        struct radv_image *image,
-                                        const VkImageSubresourceRange *subresourceRange,
-                                        struct radv_sample_locations_state *sample_locs)
+void radv_decompress_depth_stencil(struct radv_cmd_buffer *cmd_buffer,
+                                  struct radv_image *image,
+                                  const VkImageSubresourceRange *subresourceRange,
+                                  struct radv_sample_locations_state *sample_locs)
 {
+       struct radv_barrier_data barrier = {};
+
+       barrier.layout_transitions.depth_stencil_expand = 1;
+       radv_describe_layout_transition(cmd_buffer, &barrier);
+
        assert(cmd_buffer->queue_family_index == RADV_QUEUE_GENERAL);
-       radv_process_depth_image_inplace(cmd_buffer, image, subresourceRange,
-                                        sample_locs, DEPTH_DECOMPRESS);
+       radv_process_depth_stencil(cmd_buffer, image, subresourceRange,
+                                  sample_locs, DEPTH_DECOMPRESS);
 }
 
-void radv_resummarize_depth_image_inplace(struct radv_cmd_buffer *cmd_buffer,
-                                        struct radv_image *image,
-                                        const VkImageSubresourceRange *subresourceRange,
-                                        struct radv_sample_locations_state *sample_locs)
+void radv_resummarize_depth_stencil(struct radv_cmd_buffer *cmd_buffer,
+                                   struct radv_image *image,
+                                   const VkImageSubresourceRange *subresourceRange,
+                                   struct radv_sample_locations_state *sample_locs)
 {
+       struct radv_barrier_data barrier = {};
+
+       barrier.layout_transitions.depth_stencil_resummarize = 1;
+       radv_describe_layout_transition(cmd_buffer, &barrier);
+
        assert(cmd_buffer->queue_family_index == RADV_QUEUE_GENERAL);
-       radv_process_depth_image_inplace(cmd_buffer, image, subresourceRange,
-                                        sample_locs, DEPTH_RESUMMARIZE);
+       radv_process_depth_stencil(cmd_buffer, image, subresourceRange,
+                                  sample_locs, DEPTH_RESUMMARIZE);
 }