early-access version 3996

src/video_core/buffer_cache/buffer_cache.h

@@ -1753,15 +1753,25 @@ Binding BufferCache<P>::StorageBufferBinding(GPUVAddr ssbo_addr, u32 cbuf_index,
         const u32 memory_layout_size = static_cast<u32>(gpu_memory->GetMemoryLayoutSize(gpu_addr));
         return std::min(memory_layout_size, static_cast<u32>(8_MiB));
     }();
-    const std::optional<VAddr> cpu_addr = gpu_memory->GpuToCpuAddress(gpu_addr);
-    if (!cpu_addr || size == 0) {
+    // Alignment only applies to the offset of the buffer
+    const u32 alignment = runtime.GetStorageBufferAlignment();
+    const GPUVAddr aligned_gpu_addr = Common::AlignDown(gpu_addr, alignment);
+    const u32 aligned_size = static_cast<u32>(gpu_addr - aligned_gpu_addr) + size;
+
+    const std::optional<VAddr> aligned_cpu_addr = gpu_memory->GpuToCpuAddress(aligned_gpu_addr);
+    if (!aligned_cpu_addr || size == 0) {
         LOG_WARNING(HW_GPU, "Failed to find storage buffer for cbuf index {}", cbuf_index);
         return NULL_BINDING;
     }
-    const VAddr cpu_end = Common::AlignUp(*cpu_addr + size, YUZU_PAGESIZE);
+    const std::optional<VAddr> cpu_addr = gpu_memory->GpuToCpuAddress(gpu_addr);
+    ASSERT_MSG(cpu_addr, "Unaligned storage buffer address not found for cbuf index {}",
+               cbuf_index);
+    // The end address used for size calculation does not need to be aligned
+    const VAddr cpu_end = Common::AlignUp(*cpu_addr + size, Core::Memory::YUZU_PAGESIZE);
+
     const Binding binding{
-        .cpu_addr = *cpu_addr,
-        .size = is_written ? size : static_cast<u32>(cpu_end - *cpu_addr),
+        .cpu_addr = *aligned_cpu_addr,
+        .size = is_written ? aligned_size : static_cast<u32>(cpu_end - *aligned_cpu_addr),
         .buffer_id = BufferId{},
     };
     return binding;

src/video_core/renderer_opengl/gl_buffer_cache.h

@@ -191,6 +191,10 @@ public:
         return device.CanReportMemoryUsage();
     }
 
+    u32 GetStorageBufferAlignment() const {
+        return static_cast<u32>(device.GetShaderStorageBufferAlignment());
+    }
+
 private:
     static constexpr std::array PABO_LUT{
         GL_VERTEX_PROGRAM_PARAMETER_BUFFER_NV,          GL_TESS_CONTROL_PROGRAM_PARAMETER_BUFFER_NV,

src/video_core/renderer_opengl/gl_shader_cache.cpp

@@ -232,6 +232,7 @@ ShaderCache::ShaderCache(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindo
           .has_gl_bool_ref_bug = device.HasBoolRefBug(),
           .ignore_nan_fp_comparisons = true,
           .gl_max_compute_smem_size = device.GetMaxComputeSharedMemorySize(),
+          .min_ssbo_alignment = device.GetShaderStorageBufferAlignment(),
       },
       host_info{
           .support_float64 = true,
@@ -240,6 +241,7 @@ ShaderCache::ShaderCache(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindo
           .needs_demote_reorder = device.IsAmd(),
           .support_snorm_render_buffer = false,
           .support_viewport_index_layer = device.HasVertexViewportLayer(),
+          .min_ssbo_alignment = static_cast<u32>(device.GetShaderStorageBufferAlignment()),
           .support_geometry_shader_passthrough = device.HasGeometryShaderPassthrough(),
           .support_conditional_barrier = device.SupportsConditionalBarriers(),
       } {

src/video_core/renderer_vulkan/vk_buffer_cache.cpp

@@ -355,6 +355,10 @@ bool BufferCacheRuntime::CanReportMemoryUsage() const {
     return device.CanReportMemoryUsage();
 }
 
+u32 BufferCacheRuntime::GetStorageBufferAlignment() const {
+    return static_cast<u32>(device.GetStorageBufferAlignment());
+}
+
 void BufferCacheRuntime::TickFrame(VideoCommon::SlotVector<Buffer>& slot_buffers) noexcept {
     for (auto it = slot_buffers.begin(); it != slot_buffers.end(); it++) {
         it->ResetUsageTracking();

src/video_core/renderer_vulkan/vk_buffer_cache.h

@@ -91,6 +91,8 @@ public:
 
     bool CanReportMemoryUsage() const;
 
+    u32 GetStorageBufferAlignment() const;
+
     [[nodiscard]] StagingBufferRef UploadStagingBuffer(size_t size);
 
     [[nodiscard]] StagingBufferRef DownloadStagingBuffer(size_t size, bool deferred = false);

src/video_core/renderer_vulkan/vk_pipeline_cache.cpp

@@ -373,6 +373,7 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device
             driver_id == VK_DRIVER_ID_QUALCOMM_PROPRIETARY,
         .has_broken_robust =
             device.IsNvidia() && device.GetNvidiaArch() <= NvidiaArchitecture::Arch_Pascal,
+        .min_ssbo_alignment = device.GetStorageBufferAlignment(),
     };
 
     host_info = Shader::HostTranslateInfo{
@@ -383,6 +384,7 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device
             driver_id == VK_DRIVER_ID_AMD_PROPRIETARY || driver_id == VK_DRIVER_ID_AMD_OPEN_SOURCE,
         .support_snorm_render_buffer = true,
         .support_viewport_index_layer = device.IsExtShaderViewportIndexLayerSupported(),
+        .min_ssbo_alignment = static_cast<u32>(device.GetStorageBufferAlignment()),
         .support_geometry_shader_passthrough = device.IsNvGeometryShaderPassthroughSupported(),
         .support_conditional_barrier = device.SupportsConditionalBarriers(),
     };