shader_recompiler: Align SSBO offsets to meet host requirements

Co-Authored-By: Billy Laws <blaws05@gmail.com>

6 files changed, 20 insertions, 3 deletions

diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h
index 081a574e8..9202e53c7 100644
--- a/src/video_core/buffer_cache/buffer_cache.h
+++ b/src/video_core/buffer_cache/buffer_cache.h
@@ -1770,6 +1770,7 @@ template <class P>
 Binding BufferCache<P>::StorageBufferBinding(GPUVAddr ssbo_addr, u32 cbuf_index,
                                              bool is_written) const {
     const GPUVAddr gpu_addr = gpu_memory->Read<u64>(ssbo_addr);
+    const u32 alignment = runtime.GetStorageBufferAlignment();
     const auto size = [&]() {
         const bool is_nvn_cbuf = cbuf_index == 0;
         // The NVN driver buffer (index 0) is known to pack the SSBO address followed by its size.
@@ -1785,15 +1786,19 @@ 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);
+    const GPUVAddr aligned_gpu_addr = Common::AlignDown(gpu_addr, alignment);
+    const u32 aligned_size =
+        Common::AlignUp(static_cast<u32>(gpu_addr - aligned_gpu_addr) + size, alignment);
+
+    const std::optional<VAddr> cpu_addr = gpu_memory->GpuToCpuAddress(aligned_gpu_addr);
     if (!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 VAddr cpu_end = Common::AlignUp(*cpu_addr + aligned_size, Core::Memory::YUZU_PAGESIZE);
     const Binding binding{
         .cpu_addr = *cpu_addr,
-        .size = is_written ? size : static_cast<u32>(cpu_end - *cpu_addr),
+        .size = is_written ? aligned_size : static_cast<u32>(cpu_end - *cpu_addr),
         .buffer_id = BufferId{},
     };
     return binding;
diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h
index 41b746f3b..e8dbbd3a2 100644
--- a/src/video_core/renderer_opengl/gl_buffer_cache.h
+++ b/src/video_core/renderer_opengl/gl_buffer_cache.h
@@ -182,6 +182,10 @@ 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,
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index 2888e0238..69f6759e6 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -240,6 +240,7 @@ 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(),
       } {
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
index d8148e89a..976c3f6a6 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
+++ b/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::Finish() {
     scheduler.Finish();
 }
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h
index 95446c732..833dfac45 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.h
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h
@@ -75,6 +75,8 @@ public:
 
     bool CanReportMemoryUsage() const;
 
+    u32 GetStorageBufferAlignment() const;
+
     [[nodiscard]] StagingBufferRef UploadStagingBuffer(size_t size);
 
     [[nodiscard]] StagingBufferRef DownloadStagingBuffer(size_t size, bool deferred = false);
diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp
index 22bf8cc77..5e4380175 100644
--- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp
+++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp
@@ -369,6 +369,7 @@ 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(),
     };