diff options
Diffstat (limited to 'src/video_core')
19 files changed, 178 insertions, 88 deletions
diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index 90dbd352f..6d1fc3887 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/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; diff --git a/src/video_core/buffer_cache/usage_tracker.h b/src/video_core/buffer_cache/usage_tracker.h index ab05fe415..5f8688d31 100644 --- a/src/video_core/buffer_cache/usage_tracker.h +++ b/src/video_core/buffer_cache/usage_tracker.h @@ -58,7 +58,7 @@ private: void TrackPage(u64 page, u64 offset, u64 size) noexcept { const size_t offset_in_page = offset % PAGE_BYTES; const size_t first_bit = offset_in_page >> BYTES_PER_BIT_SHIFT; - const size_t num_bits = std::min(size, PAGE_BYTES) >> BYTES_PER_BIT_SHIFT; + const size_t num_bits = std::min<size_t>(size, PAGE_BYTES) >> BYTES_PER_BIT_SHIFT; const size_t mask = ~u64{0} >> (64 - num_bits); pages[page] |= (~u64{0} & mask) << first_bit; } @@ -66,7 +66,7 @@ private: bool IsPageUsed(u64 page, u64 offset, u64 size) const noexcept { const size_t offset_in_page = offset % PAGE_BYTES; const size_t first_bit = offset_in_page >> BYTES_PER_BIT_SHIFT; - const size_t num_bits = std::min(size, PAGE_BYTES) >> BYTES_PER_BIT_SHIFT; + const size_t num_bits = std::min<size_t>(size, PAGE_BYTES) >> BYTES_PER_BIT_SHIFT; const size_t mask = ~u64{0} >> (64 - num_bits); const size_t mask2 = (~u64{0} & mask) << first_bit; return (pages[page] & mask2) != 0; diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp index dfd696de6..ed188b435 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp @@ -146,8 +146,12 @@ StagingBufferMap BufferCacheRuntime::UploadStagingBuffer(size_t size) { return staging_buffer_pool.RequestUploadBuffer(size); } -StagingBufferMap BufferCacheRuntime::DownloadStagingBuffer(size_t size) { - return staging_buffer_pool.RequestDownloadBuffer(size); +StagingBufferMap BufferCacheRuntime::DownloadStagingBuffer(size_t size, bool deferred) { + return staging_buffer_pool.RequestDownloadBuffer(size, deferred); +} + +void BufferCacheRuntime::FreeDeferredStagingBuffer(StagingBufferMap& buffer) { + staging_buffer_pool.FreeDeferredStagingBuffer(buffer); } u64 BufferCacheRuntime::GetDeviceMemoryUsage() const { diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h index feccf06f9..1e8708f59 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.h +++ b/src/video_core/renderer_opengl/gl_buffer_cache.h @@ -66,7 +66,9 @@ public: [[nodiscard]] StagingBufferMap UploadStagingBuffer(size_t size); - [[nodiscard]] StagingBufferMap DownloadStagingBuffer(size_t size); + [[nodiscard]] StagingBufferMap DownloadStagingBuffer(size_t size, bool deferred = false); + + void FreeDeferredStagingBuffer(StagingBufferMap& buffer); bool CanReorderUpload(const Buffer&, std::span<const VideoCommon::BufferCopy>) { return false; @@ -191,6 +193,10 @@ 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, @@ -242,7 +248,7 @@ struct BufferCacheParams { static constexpr bool NEEDS_BIND_STORAGE_INDEX = true; static constexpr bool USE_MEMORY_MAPS = true; static constexpr bool SEPARATE_IMAGE_BUFFER_BINDINGS = true; - static constexpr bool IMPLEMENTS_ASYNC_DOWNLOADS = false; + static constexpr bool IMPLEMENTS_ASYNC_DOWNLOADS = true; // TODO: Investigate why OpenGL seems to perform worse with persistently mapped buffer uploads static constexpr bool USE_MEMORY_MAPS_FOR_UPLOADS = false; diff --git a/src/video_core/renderer_opengl/gl_device.cpp b/src/video_core/renderer_opengl/gl_device.cpp index 94258ccd0..993438a27 100644 --- a/src/video_core/renderer_opengl/gl_device.cpp +++ b/src/video_core/renderer_opengl/gl_device.cpp @@ -265,33 +265,33 @@ std::string Device::GetVendorName() const { if (vendor_name == "Intel") { // For Mesa, `Intel` is an overloaded vendor string that could mean crocus or iris. // Simply return `INTEL` for those as well as the Windows driver. - return "INTEL"; + return "Intel"; } if (vendor_name == "Intel Open Source Technology Center") { - return "I965"; + return "i965"; } if (vendor_name == "Mesa Project") { - return "I915"; + return "i915"; } if (vendor_name == "Mesa/X.org") { // This vendor string is overloaded between llvmpipe, softpipe, and virgl, so just return // MESA instead of one of those driver names. - return "MESA"; + return "Mesa"; } if (vendor_name == "AMD") { - return "RADEONSI"; + return "RadeonSI"; } if (vendor_name == "nouveau") { - return "NOUVEAU"; + return "Nouveau"; } if (vendor_name == "X.Org") { return "R600"; } if (vendor_name == "Collabora Ltd") { - return "ZINK"; + return "Zink"; } if (vendor_name == "Intel Corporation") { - return "OPENSWR"; + return "OpenSWR"; } if (vendor_name == "Microsoft Corporation") { return "D3D12"; @@ -300,7 +300,7 @@ std::string Device::GetVendorName() const { // Mesa's tegra driver reports `NVIDIA`. Only present in this list because the default // strategy would have returned `NVIDIA` here for this driver, the same result as the // proprietary driver. - return "TEGRA"; + return "Tegra"; } return vendor_name; } diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 2888e0238..26f2d0ea7 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/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(), } { diff --git a/src/video_core/renderer_opengl/gl_staging_buffer_pool.cpp b/src/video_core/renderer_opengl/gl_staging_buffer_pool.cpp index bbb06e51f..cadad6507 100644 --- a/src/video_core/renderer_opengl/gl_staging_buffer_pool.cpp +++ b/src/video_core/renderer_opengl/gl_staging_buffer_pool.cpp @@ -28,63 +28,69 @@ StagingBuffers::StagingBuffers(GLenum storage_flags_, GLenum map_flags_) StagingBuffers::~StagingBuffers() = default; -StagingBufferMap StagingBuffers::RequestMap(size_t requested_size, bool insert_fence) { +StagingBufferMap StagingBuffers::RequestMap(size_t requested_size, bool insert_fence, + bool deferred) { MICROPROFILE_SCOPE(OpenGL_BufferRequest); const size_t index = RequestBuffer(requested_size); - OGLSync* const sync = insert_fence ? &syncs[index] : nullptr; - sync_indices[index] = insert_fence ? ++current_sync_index : 0; + OGLSync* const sync = insert_fence ? &allocs[index].sync : nullptr; + allocs[index].sync_index = insert_fence ? ++current_sync_index : 0; + allocs[index].deferred = deferred; return StagingBufferMap{ - .mapped_span = std::span(maps[index], requested_size), + .mapped_span = std::span(allocs[index].map, requested_size), .sync = sync, - .buffer = buffers[index].handle, + .buffer = allocs[index].buffer.handle, + .index = index, }; } +void StagingBuffers::FreeDeferredStagingBuffer(size_t index) { + ASSERT(allocs[index].deferred); + allocs[index].deferred = false; +} + size_t StagingBuffers::RequestBuffer(size_t requested_size) { if (const std::optional<size_t> index = FindBuffer(requested_size); index) { return *index; } - - OGLBuffer& buffer = buffers.emplace_back(); - buffer.Create(); + StagingBufferAlloc alloc; + alloc.buffer.Create(); const auto next_pow2_size = Common::NextPow2(requested_size); - glNamedBufferStorage(buffer.handle, next_pow2_size, nullptr, + glNamedBufferStorage(alloc.buffer.handle, next_pow2_size, nullptr, storage_flags | GL_MAP_PERSISTENT_BIT); - maps.push_back(static_cast<u8*>(glMapNamedBufferRange(buffer.handle, 0, next_pow2_size, - map_flags | GL_MAP_PERSISTENT_BIT))); - syncs.emplace_back(); - sync_indices.emplace_back(); - sizes.push_back(next_pow2_size); - - ASSERT(syncs.size() == buffers.size() && buffers.size() == maps.size() && - maps.size() == sizes.size()); - - return buffers.size() - 1; + alloc.map = static_cast<u8*>(glMapNamedBufferRange(alloc.buffer.handle, 0, next_pow2_size, + map_flags | GL_MAP_PERSISTENT_BIT)); + alloc.size = next_pow2_size; + allocs.emplace_back(std::move(alloc)); + return allocs.size() - 1; } std::optional<size_t> StagingBuffers::FindBuffer(size_t requested_size) { size_t known_unsignaled_index = current_sync_index + 1; size_t smallest_buffer = std::numeric_limits<size_t>::max(); std::optional<size_t> found; - const size_t num_buffers = sizes.size(); + const size_t num_buffers = allocs.size(); for (size_t index = 0; index < num_buffers; ++index) { - const size_t buffer_size = sizes[index]; + StagingBufferAlloc& alloc = allocs[index]; + const size_t buffer_size = alloc.size; if (buffer_size < requested_size || buffer_size >= smallest_buffer) { continue; } - if (syncs[index].handle != 0) { - if (sync_indices[index] >= known_unsignaled_index) { + if (alloc.deferred) { + continue; + } + if (alloc.sync.handle != 0) { + if (alloc.sync_index >= known_unsignaled_index) { // This fence is later than a fence that is known to not be signaled continue; } - if (!syncs[index].IsSignaled()) { + if (!alloc.sync.IsSignaled()) { // Since this fence hasn't been signaled, it's safe to assume all later // fences haven't been signaled either - known_unsignaled_index = std::min(known_unsignaled_index, sync_indices[index]); + known_unsignaled_index = std::min(known_unsignaled_index, alloc.sync_index); continue; } - syncs[index].Release(); + alloc.sync.Release(); } smallest_buffer = buffer_size; found = index; @@ -143,8 +149,12 @@ StagingBufferMap StagingBufferPool::RequestUploadBuffer(size_t size) { return upload_buffers.RequestMap(size, true); } -StagingBufferMap StagingBufferPool::RequestDownloadBuffer(size_t size) { - return download_buffers.RequestMap(size, false); +StagingBufferMap StagingBufferPool::RequestDownloadBuffer(size_t size, bool deferred) { + return download_buffers.RequestMap(size, false, deferred); +} + +void StagingBufferPool::FreeDeferredStagingBuffer(StagingBufferMap& buffer) { + download_buffers.FreeDeferredStagingBuffer(buffer.index); } } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_staging_buffer_pool.h b/src/video_core/renderer_opengl/gl_staging_buffer_pool.h index 60f72d3a0..07a56b4d2 100644 --- a/src/video_core/renderer_opengl/gl_staging_buffer_pool.h +++ b/src/video_core/renderer_opengl/gl_staging_buffer_pool.h @@ -26,23 +26,30 @@ struct StagingBufferMap { size_t offset = 0; OGLSync* sync; GLuint buffer; + size_t index; }; struct StagingBuffers { explicit StagingBuffers(GLenum storage_flags_, GLenum map_flags_); ~StagingBuffers(); - StagingBufferMap RequestMap(size_t requested_size, bool insert_fence); + StagingBufferMap RequestMap(size_t requested_size, bool insert_fence, bool deferred = false); + + void FreeDeferredStagingBuffer(size_t index); size_t RequestBuffer(size_t requested_size); std::optional<size_t> FindBuffer(size_t requested_size); - std::vector<OGLSync> syncs; - std::vector<OGLBuffer> buffers; - std::vector<u8*> maps; - std::vector<size_t> sizes; - std::vector<size_t> sync_indices; + struct StagingBufferAlloc { + OGLSync sync; + OGLBuffer buffer; + u8* map; + size_t size; + size_t sync_index; + bool deferred; + }; + std::vector<StagingBufferAlloc> allocs; GLenum storage_flags; GLenum map_flags; size_t current_sync_index = 0; @@ -85,7 +92,8 @@ public: ~StagingBufferPool() = default; StagingBufferMap RequestUploadBuffer(size_t size); - StagingBufferMap RequestDownloadBuffer(size_t size); + StagingBufferMap RequestDownloadBuffer(size_t size, bool deferred = false); + void FreeDeferredStagingBuffer(StagingBufferMap& buffer); private: StagingBuffers upload_buffers{GL_MAP_WRITE_BIT, GL_MAP_WRITE_BIT | GL_MAP_FLUSH_EXPLICIT_BIT}; diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index 512eef575..66a5ca03e 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -557,8 +557,12 @@ StagingBufferMap TextureCacheRuntime::UploadStagingBuffer(size_t size) { return staging_buffer_pool.RequestUploadBuffer(size); } -StagingBufferMap TextureCacheRuntime::DownloadStagingBuffer(size_t size) { - return staging_buffer_pool.RequestDownloadBuffer(size); +StagingBufferMap TextureCacheRuntime::DownloadStagingBuffer(size_t size, bool deferred) { + return staging_buffer_pool.RequestDownloadBuffer(size, deferred); +} + +void TextureCacheRuntime::FreeDeferredStagingBuffer(StagingBufferMap& buffer) { + staging_buffer_pool.FreeDeferredStagingBuffer(buffer); } u64 TextureCacheRuntime::GetDeviceMemoryUsage() const { diff --git a/src/video_core/renderer_opengl/gl_texture_cache.h b/src/video_core/renderer_opengl/gl_texture_cache.h index e71b87e99..34870c81f 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.h +++ b/src/video_core/renderer_opengl/gl_texture_cache.h @@ -74,7 +74,9 @@ public: StagingBufferMap UploadStagingBuffer(size_t size); - StagingBufferMap DownloadStagingBuffer(size_t size); + StagingBufferMap DownloadStagingBuffer(size_t size, bool deferred = false); + + void FreeDeferredStagingBuffer(StagingBufferMap& buffer); u64 GetDeviceLocalMemory() const { return device_access_memory; @@ -359,7 +361,7 @@ struct TextureCacheParams { static constexpr bool FRAMEBUFFER_BLITS = true; static constexpr bool HAS_EMULATED_COPIES = true; static constexpr bool HAS_DEVICE_MEMORY_INFO = true; - static constexpr bool IMPLEMENTS_ASYNC_DOWNLOADS = false; + static constexpr bool IMPLEMENTS_ASYNC_DOWNLOADS = true; using Runtime = OpenGL::TextureCacheRuntime; using Image = OpenGL::Image; @@ -367,7 +369,7 @@ struct TextureCacheParams { using ImageView = OpenGL::ImageView; using Sampler = OpenGL::Sampler; using Framebuffer = OpenGL::Framebuffer; - using AsyncBuffer = u32; + using AsyncBuffer = OpenGL::StagingBufferMap; using BufferType = GLuint; }; diff --git a/src/video_core/renderer_vulkan/vk_blit_screen.cpp b/src/video_core/renderer_vulkan/vk_blit_screen.cpp index 66483a900..5e461fbd0 100644 --- a/src/video_core/renderer_vulkan/vk_blit_screen.cpp +++ b/src/video_core/renderer_vulkan/vk_blit_screen.cpp @@ -1211,7 +1211,7 @@ void BlitScreen::CreateRawImages(const Tegra::FramebufferConfig& framebuffer) { aa_framebuffer = CreateFramebuffer(*aa_image_view, size, aa_renderpass); return; } - aa_renderpass = CreateRenderPassImpl(GetFormat(framebuffer)); + aa_renderpass = CreateRenderPassImpl(VK_FORMAT_R16G16B16A16_SFLOAT); aa_framebuffer = CreateFramebuffer(*aa_image_view, size, aa_renderpass); const std::array<VkPipelineShaderStageCreateInfo, 2> fxaa_shader_stages{{ diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp index 7691cc2ba..5958f52f7 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::TickFrame(VideoCommon::SlotVector<Buffer>& slot_buffers) noexcept { for (auto it = slot_buffers.begin(); it != slot_buffers.end(); it++) { it->ResetUsageTracking(); diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h index 4416a902f..0b3fbd6d0 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.h +++ b/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); diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp index 89b455bff..2a13b2a72 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp +++ b/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(), }; diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp index e0ab1eaac..07222e603 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp +++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp @@ -892,10 +892,6 @@ void RasterizerVulkan::UpdateDynamicStates() { UpdateFrontFace(regs); UpdateStencilOp(regs); - if (device.IsExtVertexInputDynamicStateSupported()) { - UpdateVertexInput(regs); - } - if (state_tracker.TouchStateEnable()) { UpdateDepthBoundsTestEnable(regs); UpdateDepthTestEnable(regs); @@ -918,6 +914,9 @@ void RasterizerVulkan::UpdateDynamicStates() { UpdateBlending(regs); } } + if (device.IsExtVertexInputDynamicStateSupported()) { + UpdateVertexInput(regs); + } } void RasterizerVulkan::HandleTransformFeedback() { diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index 5dbec2e62..38b1619df 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -1439,7 +1439,7 @@ void Image::UploadMemory(const StagingBufferRef& map, std::span<const BufferImag UploadMemory(map.buffer, map.offset, copies); } -void Image::DownloadMemory(VkBuffer buffer, VkDeviceSize offset, +void Image::DownloadMemory(VkBuffer buffer, size_t offset, std::span<const VideoCommon::BufferImageCopy> copies) { std::array buffer_handles{ buffer, @@ -1450,7 +1450,7 @@ void Image::DownloadMemory(VkBuffer buffer, VkDeviceSize offset, DownloadMemory(buffer_handles, buffer_offsets, copies); } -void Image::DownloadMemory(std::span<VkBuffer> buffers_span, std::span<VkDeviceSize> offsets_span, +void Image::DownloadMemory(std::span<VkBuffer> buffers_span, std::span<size_t> offsets_span, std::span<const VideoCommon::BufferImageCopy> copies) { const bool is_rescaled = True(flags & ImageFlagBits::Rescaled); if (is_rescaled) { @@ -1530,7 +1530,7 @@ void Image::DownloadMemory(const StagingBufferRef& map, std::span<const BufferIm map.buffer, }; std::array offsets{ - map.offset, + static_cast<size_t>(map.offset), }; DownloadMemory(buffers, offsets, copies); } diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.h b/src/video_core/renderer_vulkan/vk_texture_cache.h index edf5d7635..0dbde65d6 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.h +++ b/src/video_core/renderer_vulkan/vk_texture_cache.h @@ -147,10 +147,10 @@ public: void UploadMemory(const StagingBufferRef& map, std::span<const VideoCommon::BufferImageCopy> copies); - void DownloadMemory(VkBuffer buffer, VkDeviceSize offset, + void DownloadMemory(VkBuffer buffer, size_t offset, std::span<const VideoCommon::BufferImageCopy> copies); - void DownloadMemory(std::span<VkBuffer> buffers, std::span<VkDeviceSize> offsets, + void DownloadMemory(std::span<VkBuffer> buffers, std::span<size_t> offsets, std::span<const VideoCommon::BufferImageCopy> copies); void DownloadMemory(const StagingBufferRef& map, diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index d575c57ca..dade38b18 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -995,7 +995,7 @@ void TextureCache<P>::DownloadImageIntoBuffer(typename TextureCache<P>::Image* i buffer, download_map.buffer, }; - std::array<u64, 2> buffer_offsets{ + std::array<size_t, 2> buffer_offsets{ buffer_offset, download_map.offset, }; diff --git a/src/video_core/vulkan_common/vulkan_device.cpp b/src/video_core/vulkan_common/vulkan_device.cpp index fde36a49c..1fda0042d 100644 --- a/src/video_core/vulkan_common/vulkan_device.cpp +++ b/src/video_core/vulkan_common/vulkan_device.cpp @@ -519,10 +519,6 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR LOG_WARNING(Render_Vulkan, "ARM drivers have broken VK_EXT_extended_dynamic_state"); RemoveExtensionFeature(extensions.extended_dynamic_state, features.extended_dynamic_state, VK_EXT_EXTENDED_DYNAMIC_STATE_EXTENSION_NAME); - - LOG_WARNING(Render_Vulkan, "ARM drivers have broken VK_EXT_extended_dynamic_state2"); - RemoveExtensionFeature(extensions.extended_dynamic_state2, features.extended_dynamic_state2, - VK_EXT_EXTENDED_DYNAMIC_STATE_2_EXTENSION_NAME); } if (is_nvidia) { @@ -611,17 +607,12 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR } } if (extensions.vertex_input_dynamic_state && is_qualcomm) { - const u32 version = (properties.properties.driverVersion << 3) >> 3; - if (version >= VK_MAKE_API_VERSION(0, 0, 676, 0) && - version < VK_MAKE_API_VERSION(0, 0, 680, 0)) { - // Qualcomm Adreno 7xx drivers do not properly support vertex_input_dynamic_state. - LOG_WARNING( - Render_Vulkan, - "Qualcomm Adreno 7xx drivers have broken VK_EXT_vertex_input_dynamic_state"); - RemoveExtensionFeature(extensions.vertex_input_dynamic_state, - features.vertex_input_dynamic_state, - VK_EXT_VERTEX_INPUT_DYNAMIC_STATE_EXTENSION_NAME); - } + // Qualcomm drivers do not properly support vertex_input_dynamic_state. + LOG_WARNING(Render_Vulkan, + "Qualcomm drivers have broken VK_EXT_vertex_input_dynamic_state"); + RemoveExtensionFeature(extensions.vertex_input_dynamic_state, + features.vertex_input_dynamic_state, + VK_EXT_VERTEX_INPUT_DYNAMIC_STATE_EXTENSION_NAME); } sets_per_pool = 64; @@ -704,6 +695,22 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR std::min(properties.properties.limits.maxVertexInputBindings, 16U); } + if (!extensions.extended_dynamic_state && extensions.extended_dynamic_state2) { + LOG_INFO(Render_Vulkan, + "Removing extendedDynamicState2 due to missing extendedDynamicState"); + RemoveExtensionFeature(extensions.extended_dynamic_state2, features.extended_dynamic_state2, + VK_EXT_EXTENDED_DYNAMIC_STATE_2_EXTENSION_NAME); + } + + if (!extensions.extended_dynamic_state2 && extensions.extended_dynamic_state3) { + LOG_INFO(Render_Vulkan, + "Removing extendedDynamicState3 due to missing extendedDynamicState2"); + RemoveExtensionFeature(extensions.extended_dynamic_state3, features.extended_dynamic_state3, + VK_EXT_EXTENDED_DYNAMIC_STATE_3_EXTENSION_NAME); + dynamic_state3_blending = false; + dynamic_state3_enables = false; + } + logical = vk::Device::Create(physical, queue_cis, ExtensionListForVulkan(loaded_extensions), first_next, dld); @@ -847,11 +854,41 @@ std::string Device::GetDriverName() const { case VK_DRIVER_ID_NVIDIA_PROPRIETARY: return "NVIDIA"; case VK_DRIVER_ID_INTEL_PROPRIETARY_WINDOWS: - return "INTEL"; + return "Intel"; case VK_DRIVER_ID_INTEL_OPEN_SOURCE_MESA: return "ANV"; + case VK_DRIVER_ID_IMAGINATION_PROPRIETARY: + return "PowerVR"; + case VK_DRIVER_ID_QUALCOMM_PROPRIETARY: + return "Qualcomm"; + case VK_DRIVER_ID_ARM_PROPRIETARY: + return "Mali"; + case VK_DRIVER_ID_GOOGLE_SWIFTSHADER: + return "SwiftShader"; + case VK_DRIVER_ID_BROADCOM_PROPRIETARY: + return "Broadcom"; case VK_DRIVER_ID_MESA_LLVMPIPE: - return "LAVAPIPE"; + return "Lavapipe"; + case VK_DRIVER_ID_MOLTENVK: + return "MoltenVK"; + case VK_DRIVER_ID_VERISILICON_PROPRIETARY: + return "Vivante"; + case VK_DRIVER_ID_MESA_TURNIP: + return "Turnip"; + case VK_DRIVER_ID_MESA_V3DV: + return "V3DV"; + case VK_DRIVER_ID_MESA_PANVK: + return "PanVK"; + case VK_DRIVER_ID_MESA_VENUS: + return "Venus"; + case VK_DRIVER_ID_MESA_DOZEN: + return "Dozen"; + case VK_DRIVER_ID_MESA_NVK: + return "NVK"; + case VK_DRIVER_ID_IMAGINATION_OPEN_SOURCE_MESA: + return "PVR"; + // case VK_DRIVER_ID_MESA_AGXV: + // return "Asahi"; default: return properties.driver.driverName; } |