Rasterizer: Refactor inlineToMemory.

src/video_core/buffer_cache/buffer_cache.h

@@ -1661,6 +1661,7 @@ bool BufferCache<P>::InlineMemory(VAddr dest_address, size_t copy_size,
 
     const IntervalType subtract_interval{dest_address, dest_address + copy_size};
     ClearDownload(subtract_interval);
+    common_ranges.subtract(subtract_interval);
 
     BufferId buffer_id = FindBuffer(dest_address, static_cast<u32>(copy_size));
     auto& buffer = slot_buffers[buffer_id];
@@ -1677,7 +1678,7 @@ bool BufferCache<P>::InlineMemory(VAddr dest_address, size_t copy_size,
         std::memcpy(src_pointer, inlined_buffer.data(), copy_size);
         runtime.CopyBuffer(buffer, upload_staging.buffer, copies);
     } else {
-        buffer.ImmediateUpload(buffer.Offset(dest_address), inlined_buffer);
+        buffer.ImmediateUpload(buffer.Offset(dest_address), inlined_buffer.first(copy_size));
     }
 
     return true;

src/video_core/engines/engine_upload.cpp

@@ -37,7 +37,7 @@ void State::ProcessData(const u32 data, const bool is_last_call) {
     }
     const GPUVAddr address{regs.dest.Address()};
     if (is_linear) {
-        rasterizer->AccelerateInline2Memory(address, copy_size, inner_buffer);
+        rasterizer->AccelerateInlineToMemory(address, copy_size, inner_buffer);
     } else {
         UNIMPLEMENTED_IF(regs.dest.z != 0);
         UNIMPLEMENTED_IF(regs.dest.depth != 1);

src/video_core/engines/maxwell_3d.h

@@ -1557,7 +1557,7 @@ private:
 
     static constexpr u32 null_cb_data = 0xFFFFFFFF;
     struct CBDataState {
-        static constexpr size_t inline_size = 0x8000;
+        static constexpr size_t inline_size = 0x4000;
         std::array<std::array<u32, inline_size>, 16> buffer;
         u32 current{null_cb_data};
         u32 id{null_cb_data};

src/video_core/gpu.cpp

@@ -504,14 +504,12 @@ struct GPU::Impl {
         case BufferMethods::SemaphoreAddressLow:
         case BufferMethods::SemaphoreSequence:
             break;
-        case BufferMethods::UnkCacheFlush: {
+        case BufferMethods::UnkCacheFlush:
             rasterizer->SyncGuestHost();
             break;
-        }
-        case BufferMethods::WrcacheFlush: {
+        case BufferMethods::WrcacheFlush:
             rasterizer->SignalReference();
             break;
-        }
         case BufferMethods::FenceValue:
             break;
         case BufferMethods::RefCnt:

src/video_core/rasterizer_interface.h

@@ -123,7 +123,7 @@ public:
 
     [[nodiscard]] virtual Tegra::Engines::AccelerateDMAInterface& AccessAccelerateDMA() = 0;
 
-    virtual void AccelerateInline2Memory(GPUVAddr address, size_t copy_size,
+    virtual void AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
                                           std::span<u8> memory) = 0;
 
     /// Attempt to use a faster method to display the framebuffer to screen

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -484,7 +484,7 @@ Tegra::Engines::AccelerateDMAInterface& RasterizerOpenGL::AccessAccelerateDMA()
     return accelerate_dma;
 }
 
-void RasterizerOpenGL::AccelerateInline2Memory(GPUVAddr address, size_t copy_size,
+void RasterizerOpenGL::AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
                                                 std::span<u8> memory) {
     auto cpu_addr = gpu_memory.GpuToCpuAddress(address);
     if (!cpu_addr) [[unlikely]] {

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -106,7 +106,8 @@ public:
                                const Tegra::Engines::Fermi2D::Surface& dst,
                                const Tegra::Engines::Fermi2D::Config& copy_config) override;
     Tegra::Engines::AccelerateDMAInterface& AccessAccelerateDMA() override;
-    void AccelerateInline2Memory(GPUVAddr address, size_t copy_size, std::span<u8> memory) override;
+    void AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
+                                  std::span<u8> memory) override;
     bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
                            u32 pixel_stride) override;
     void LoadDiskResources(u64 title_id, std::stop_token stop_loading,

src/video_core/renderer_vulkan/vk_rasterizer.cpp

@@ -548,7 +548,7 @@ Tegra::Engines::AccelerateDMAInterface& RasterizerVulkan::AccessAccelerateDMA()
     return accelerate_dma;
 }
 
-void RasterizerVulkan::AccelerateInline2Memory(GPUVAddr address, size_t copy_size,
+void RasterizerVulkan::AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
                                                 std::span<u8> memory) {
     auto cpu_addr = gpu_memory.GpuToCpuAddress(address);
     if (!cpu_addr) [[unlikely]] {

src/video_core/renderer_vulkan/vk_rasterizer.h

@@ -99,7 +99,8 @@ public:
                                const Tegra::Engines::Fermi2D::Surface& dst,
                                const Tegra::Engines::Fermi2D::Config& copy_config) override;
     Tegra::Engines::AccelerateDMAInterface& AccessAccelerateDMA() override;
-    void AccelerateInline2Memory(GPUVAddr address, size_t copy_size, std::span<u8> memory) override;
+    void AccelerateInlineToMemory(GPUVAddr address, size_t copy_size,
+                                  std::span<u8> memory) override;
     bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
                            u32 pixel_stride) override;
     void LoadDiskResources(u64 title_id, std::stop_token stop_loading,