diff options
Diffstat (limited to '')
34 files changed, 501 insertions, 215 deletions
diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index 2f281b370..45977d578 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -715,6 +715,8 @@ void BufferCache<P>::BindHostIndexBuffer() { template <class P> void BufferCache<P>::BindHostVertexBuffers() { + HostBindings<typename P::Buffer> host_bindings; + bool any_valid{false}; auto& flags = maxwell3d->dirty.flags; for (u32 index = 0; index < NUM_VERTEX_BUFFERS; ++index) { const Binding& binding = channel_state->vertex_buffers[index]; @@ -726,9 +728,28 @@ void BufferCache<P>::BindHostVertexBuffers() { } flags[Dirty::VertexBuffer0 + index] = false; - const u32 stride = maxwell3d->regs.vertex_streams[index].stride; - const u32 offset = buffer.Offset(binding.cpu_addr); - runtime.BindVertexBuffer(index, buffer, offset, binding.size, stride); + host_bindings.min_index = std::min(host_bindings.min_index, index); + host_bindings.max_index = std::max(host_bindings.max_index, index); + any_valid = true; + } + + if (any_valid) { + host_bindings.max_index++; + for (u32 index = host_bindings.min_index; index < host_bindings.max_index; index++) { + flags[Dirty::VertexBuffer0 + index] = false; + + const Binding& binding = channel_state->vertex_buffers[index]; + Buffer& buffer = slot_buffers[binding.buffer_id]; + + const u32 stride = maxwell3d->regs.vertex_streams[index].stride; + const u32 offset = buffer.Offset(binding.cpu_addr); + + host_bindings.buffers.push_back(&buffer); + host_bindings.offsets.push_back(offset); + host_bindings.sizes.push_back(binding.size); + host_bindings.strides.push_back(stride); + } + runtime.BindVertexBuffers(host_bindings); } } @@ -882,15 +903,25 @@ void BufferCache<P>::BindHostTransformFeedbackBuffers() { if (maxwell3d->regs.transform_feedback_enabled == 0) { return; } + HostBindings<typename P::Buffer> host_bindings; for (u32 index = 0; index < NUM_TRANSFORM_FEEDBACK_BUFFERS; ++index) { const Binding& binding = channel_state->transform_feedback_buffers[index]; + if (maxwell3d->regs.transform_feedback.controls[index].varying_count == 0 && + maxwell3d->regs.transform_feedback.controls[index].stride == 0) { + break; + } Buffer& buffer = slot_buffers[binding.buffer_id]; TouchBuffer(buffer, binding.buffer_id); const u32 size = binding.size; SynchronizeBuffer(buffer, binding.cpu_addr, size); const u32 offset = buffer.Offset(binding.cpu_addr); - runtime.BindTransformFeedbackBuffer(index, buffer, offset, size); + host_bindings.buffers.push_back(&buffer); + host_bindings.offsets.push_back(offset); + host_bindings.sizes.push_back(binding.size); + } + if (host_bindings.buffers.size() > 0) { + runtime.BindTransformFeedbackBuffers(host_bindings); } } @@ -1616,6 +1647,8 @@ void BufferCache<P>::DownloadBufferMemory(Buffer& buffer, VAddr cpu_addr, u64 si template <class P> void BufferCache<P>::DeleteBuffer(BufferId buffer_id, bool do_not_mark) { + bool dirty_index{false}; + boost::container::small_vector<u64, NUM_VERTEX_BUFFERS> dirty_vertex_buffers; const auto scalar_replace = [buffer_id](Binding& binding) { if (binding.buffer_id == buffer_id) { binding.buffer_id = BufferId{}; @@ -1624,8 +1657,19 @@ void BufferCache<P>::DeleteBuffer(BufferId buffer_id, bool do_not_mark) { const auto replace = [scalar_replace](std::span<Binding> bindings) { std::ranges::for_each(bindings, scalar_replace); }; - scalar_replace(channel_state->index_buffer); - replace(channel_state->vertex_buffers); + + if (channel_state->index_buffer.buffer_id == buffer_id) { + channel_state->index_buffer.buffer_id = BufferId{}; + dirty_index = true; + } + + for (u32 index = 0; index < channel_state->vertex_buffers.size(); index++) { + auto& binding = channel_state->vertex_buffers[index]; + if (binding.buffer_id == buffer_id) { + binding.buffer_id = BufferId{}; + dirty_vertex_buffers.push_back(index); + } + } std::ranges::for_each(channel_state->uniform_buffers, replace); std::ranges::for_each(channel_state->storage_buffers, replace); replace(channel_state->transform_feedback_buffers); @@ -1642,20 +1686,21 @@ void BufferCache<P>::DeleteBuffer(BufferId buffer_id, bool do_not_mark) { delayed_destruction_ring.Push(std::move(slot_buffers[buffer_id])); slot_buffers.erase(buffer_id); - NotifyBufferDeletion(); -} - -template <class P> -void BufferCache<P>::NotifyBufferDeletion() { if constexpr (HAS_PERSISTENT_UNIFORM_BUFFER_BINDINGS) { channel_state->dirty_uniform_buffers.fill(~u32{0}); channel_state->uniform_buffer_binding_sizes.fill({}); } + auto& flags = maxwell3d->dirty.flags; - flags[Dirty::IndexBuffer] = true; - flags[Dirty::VertexBuffers] = true; - for (u32 index = 0; index < NUM_VERTEX_BUFFERS; ++index) { - flags[Dirty::VertexBuffer0 + index] = true; + if (dirty_index) { + flags[Dirty::IndexBuffer] = true; + } + + if (dirty_vertex_buffers.size() > 0) { + flags[Dirty::VertexBuffers] = true; + for (auto index : dirty_vertex_buffers) { + flags[Dirty::VertexBuffer0 + index] = true; + } } channel_state->has_deleted_buffers = true; } diff --git a/src/video_core/buffer_cache/buffer_cache_base.h b/src/video_core/buffer_cache/buffer_cache_base.h index 60a1f285e..63a120f7a 100644 --- a/src/video_core/buffer_cache/buffer_cache_base.h +++ b/src/video_core/buffer_cache/buffer_cache_base.h @@ -105,6 +105,16 @@ static constexpr Binding NULL_BINDING{ .buffer_id = NULL_BUFFER_ID, }; +template <typename Buffer> +struct HostBindings { + boost::container::small_vector<Buffer*, NUM_VERTEX_BUFFERS> buffers; + boost::container::small_vector<u64, NUM_VERTEX_BUFFERS> offsets; + boost::container::small_vector<u64, NUM_VERTEX_BUFFERS> sizes; + boost::container::small_vector<u64, NUM_VERTEX_BUFFERS> strides; + u32 min_index{NUM_VERTEX_BUFFERS}; + u32 max_index{0}; +}; + class BufferCacheChannelInfo : public ChannelInfo { public: BufferCacheChannelInfo() = delete; @@ -519,8 +529,6 @@ private: void DeleteBuffer(BufferId buffer_id, bool do_not_mark = false); - void NotifyBufferDeletion(); - [[nodiscard]] Binding StorageBufferBinding(GPUVAddr ssbo_addr, u32 cbuf_index, bool is_written) const; diff --git a/src/video_core/engines/draw_manager.cpp b/src/video_core/engines/draw_manager.cpp index 0e94c521a..f34090791 100644 --- a/src/video_core/engines/draw_manager.cpp +++ b/src/video_core/engines/draw_manager.cpp @@ -1,6 +1,7 @@ // SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project // SPDX-License-Identifier: GPL-2.0-or-later +#include "common/settings.h" #include "video_core/dirty_flags.h" #include "video_core/engines/draw_manager.h" #include "video_core/rasterizer_interface.h" @@ -195,8 +196,12 @@ void DrawManager::DrawTexture() { if (lower_left) { draw_texture_state.dst_y0 -= dst_height; } - draw_texture_state.dst_x1 = draw_texture_state.dst_x0 + dst_width; - draw_texture_state.dst_y1 = draw_texture_state.dst_y0 + dst_height; + draw_texture_state.dst_x1 = + draw_texture_state.dst_x0 + + static_cast<f32>(Settings::values.resolution_info.ScaleUp(static_cast<u32>(dst_width))); + draw_texture_state.dst_y1 = + draw_texture_state.dst_y0 + + static_cast<f32>(Settings::values.resolution_info.ScaleUp(static_cast<u32>(dst_height))); draw_texture_state.src_x0 = static_cast<float>(regs.draw_texture.src_x0) / 4096.f; draw_texture_state.src_y0 = static_cast<float>(regs.draw_texture.src_y0) / 4096.f; draw_texture_state.src_x1 = @@ -207,7 +212,6 @@ void DrawManager::DrawTexture() { draw_texture_state.src_y0; draw_texture_state.src_sampler = regs.draw_texture.src_sampler; draw_texture_state.src_texture = regs.draw_texture.src_texture; - maxwell3d->rasterizer->DrawTexture(); } diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp index 7b2cde7a7..b2f7e160a 100644 --- a/src/video_core/memory_manager.cpp +++ b/src/video_core/memory_manager.cpp @@ -111,7 +111,7 @@ GPUVAddr MemoryManager::PageTableOp(GPUVAddr gpu_addr, [[maybe_unused]] VAddr cp [[maybe_unused]] const auto current_entry_type = GetEntry<false>(current_gpu_addr); SetEntry<false>(current_gpu_addr, entry_type); if (current_entry_type != entry_type) { - rasterizer->ModifyGPUMemory(unique_identifier, gpu_addr, page_size); + rasterizer->ModifyGPUMemory(unique_identifier, current_gpu_addr, page_size); } if constexpr (entry_type == EntryType::Mapped) { const VAddr current_cpu_addr = cpu_addr + offset; @@ -134,7 +134,7 @@ GPUVAddr MemoryManager::BigPageTableOp(GPUVAddr gpu_addr, [[maybe_unused]] VAddr [[maybe_unused]] const auto current_entry_type = GetEntry<true>(current_gpu_addr); SetEntry<true>(current_gpu_addr, entry_type); if (current_entry_type != entry_type) { - rasterizer->ModifyGPUMemory(unique_identifier, gpu_addr, big_page_size); + rasterizer->ModifyGPUMemory(unique_identifier, current_gpu_addr, big_page_size); } if constexpr (entry_type == EntryType::Mapped) { const VAddr current_cpu_addr = cpu_addr + offset; diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp index c419714d4..38d553d3c 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp @@ -232,6 +232,15 @@ void BufferCacheRuntime::BindVertexBuffer(u32 index, Buffer& buffer, u32 offset, } } +void BufferCacheRuntime::BindVertexBuffers(VideoCommon::HostBindings<Buffer>& bindings) { + for (u32 index = 0; index < bindings.buffers.size(); ++index) { + BindVertexBuffer(bindings.min_index + index, *bindings.buffers[index], + static_cast<u32>(bindings.offsets[index]), + static_cast<u32>(bindings.sizes[index]), + static_cast<u32>(bindings.strides[index])); + } +} + void BufferCacheRuntime::BindUniformBuffer(size_t stage, u32 binding_index, Buffer& buffer, u32 offset, u32 size) { if (use_assembly_shaders) { @@ -320,6 +329,14 @@ void BufferCacheRuntime::BindTransformFeedbackBuffer(u32 index, Buffer& buffer, static_cast<GLintptr>(offset), static_cast<GLsizeiptr>(size)); } +void BufferCacheRuntime::BindTransformFeedbackBuffers(VideoCommon::HostBindings<Buffer>& bindings) { + for (u32 index = 0; index < bindings.buffers.size(); ++index) { + glBindBufferRange(GL_TRANSFORM_FEEDBACK_BUFFER, index, bindings.buffers[index]->Handle(), + static_cast<GLintptr>(bindings.offsets[index]), + static_cast<GLsizeiptr>(bindings.sizes[index])); + } +} + void BufferCacheRuntime::BindTextureBuffer(Buffer& buffer, u32 offset, u32 size, PixelFormat format) { *texture_handles++ = buffer.View(offset, size, format); diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h index a24991585..41b746f3b 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.h +++ b/src/video_core/renderer_opengl/gl_buffer_cache.h @@ -7,7 +7,7 @@ #include <span> #include "common/common_types.h" -#include "video_core/buffer_cache/buffer_cache.h" +#include "video_core/buffer_cache/buffer_cache_base.h" #include "video_core/buffer_cache/memory_tracker_base.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_opengl/gl_device.h" @@ -88,6 +88,8 @@ public: void BindVertexBuffer(u32 index, Buffer& buffer, u32 offset, u32 size, u32 stride); + void BindVertexBuffers(VideoCommon::HostBindings<Buffer>& bindings); + void BindUniformBuffer(size_t stage, u32 binding_index, Buffer& buffer, u32 offset, u32 size); void BindComputeUniformBuffer(u32 binding_index, Buffer& buffer, u32 offset, u32 size); @@ -100,6 +102,8 @@ public: void BindTransformFeedbackBuffer(u32 index, Buffer& buffer, u32 offset, u32 size); + void BindTransformFeedbackBuffers(VideoCommon::HostBindings<Buffer>& bindings); + void BindTextureBuffer(Buffer& buffer, u32 offset, u32 size, VideoCore::Surface::PixelFormat format); diff --git a/src/video_core/renderer_opengl/gl_compute_pipeline.cpp b/src/video_core/renderer_opengl/gl_compute_pipeline.cpp index 1a0cea9b7..3151c0db8 100644 --- a/src/video_core/renderer_opengl/gl_compute_pipeline.cpp +++ b/src/video_core/renderer_opengl/gl_compute_pipeline.cpp @@ -87,7 +87,8 @@ void ComputePipeline::Configure() { texture_cache.SynchronizeComputeDescriptors(); boost::container::static_vector<VideoCommon::ImageViewInOut, MAX_TEXTURES + MAX_IMAGES> views; - std::array<GLuint, MAX_TEXTURES> samplers; + boost::container::static_vector<VideoCommon::SamplerId, MAX_TEXTURES> samplers; + std::array<GLuint, MAX_TEXTURES> gl_samplers; std::array<GLuint, MAX_TEXTURES> textures; std::array<GLuint, MAX_IMAGES> images; GLsizei sampler_binding{}; @@ -131,7 +132,6 @@ void ComputePipeline::Configure() { for (u32 index = 0; index < desc.count; ++index) { const auto handle{read_handle(desc, index)}; views.push_back({handle.first}); - samplers[sampler_binding++] = 0; } } for (const auto& desc : info.image_buffer_descriptors) { @@ -142,8 +142,8 @@ void ComputePipeline::Configure() { const auto handle{read_handle(desc, index)}; views.push_back({handle.first}); - Sampler* const sampler = texture_cache.GetComputeSampler(handle.second); - samplers[sampler_binding++] = sampler->Handle(); + VideoCommon::SamplerId sampler = texture_cache.GetComputeSamplerId(handle.second); + samplers.push_back(sampler); } } for (const auto& desc : info.image_descriptors) { @@ -186,10 +186,17 @@ void ComputePipeline::Configure() { const VideoCommon::ImageViewInOut* views_it{views.data() + num_texture_buffers + num_image_buffers}; + const VideoCommon::SamplerId* samplers_it{samplers.data()}; texture_binding += num_texture_buffers; image_binding += num_image_buffers; u32 texture_scaling_mask{}; + + for (const auto& desc : info.texture_buffer_descriptors) { + for (u32 index = 0; index < desc.count; ++index) { + gl_samplers[sampler_binding++] = 0; + } + } for (const auto& desc : info.texture_descriptors) { for (u32 index = 0; index < desc.count; ++index) { ImageView& image_view{texture_cache.GetImageView((views_it++)->id)}; @@ -198,6 +205,12 @@ void ComputePipeline::Configure() { texture_scaling_mask |= 1u << texture_binding; } ++texture_binding; + + const Sampler& sampler{texture_cache.GetSampler(*(samplers_it++))}; + const bool use_fallback_sampler{sampler.HasAddedAnisotropy() && + !image_view.SupportsAnisotropy()}; + gl_samplers[sampler_binding++] = + use_fallback_sampler ? sampler.HandleWithDefaultAnisotropy() : sampler.Handle(); } } u32 image_scaling_mask{}; @@ -228,7 +241,7 @@ void ComputePipeline::Configure() { if (texture_binding != 0) { ASSERT(texture_binding == sampler_binding); glBindTextures(0, texture_binding, textures.data()); - glBindSamplers(0, sampler_binding, samplers.data()); + glBindSamplers(0, sampler_binding, gl_samplers.data()); } if (image_binding != 0) { glBindImageTextures(0, image_binding, images.data()); diff --git a/src/video_core/renderer_opengl/gl_device.cpp b/src/video_core/renderer_opengl/gl_device.cpp index 400c21981..03d234f2f 100644 --- a/src/video_core/renderer_opengl/gl_device.cpp +++ b/src/video_core/renderer_opengl/gl_device.cpp @@ -201,6 +201,7 @@ Device::Device(Core::Frontend::EmuWindow& emu_window) { use_asynchronous_shaders = Settings::values.use_asynchronous_shaders.GetValue() && !(is_amd || (is_intel && !is_linux)) && !strict_context_required; use_driver_cache = is_nvidia; + supports_conditional_barriers = !is_intel; LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", has_variable_aoffi); LOG_INFO(Render_OpenGL, "Renderer_ComponentIndexingBug: {}", has_component_indexing_bug); diff --git a/src/video_core/renderer_opengl/gl_device.h b/src/video_core/renderer_opengl/gl_device.h index cc0b95f1a..ad27264e5 100644 --- a/src/video_core/renderer_opengl/gl_device.h +++ b/src/video_core/renderer_opengl/gl_device.h @@ -188,6 +188,10 @@ public: return strict_context_required; } + bool SupportsConditionalBarriers() const { + return supports_conditional_barriers; + } + private: static bool TestVariableAoffi(); static bool TestPreciseBug(); @@ -233,6 +237,7 @@ private: bool has_bool_ref_bug{}; bool can_report_memory{}; bool strict_context_required{}; + bool supports_conditional_barriers{}; std::string vendor_name; }; diff --git a/src/video_core/renderer_opengl/gl_graphics_pipeline.cpp b/src/video_core/renderer_opengl/gl_graphics_pipeline.cpp index 89000d6e0..c58f760b8 100644 --- a/src/video_core/renderer_opengl/gl_graphics_pipeline.cpp +++ b/src/video_core/renderer_opengl/gl_graphics_pipeline.cpp @@ -275,9 +275,9 @@ GraphicsPipeline::GraphicsPipeline(const Device& device, TextureCache& texture_c template <typename Spec> void GraphicsPipeline::ConfigureImpl(bool is_indexed) { std::array<VideoCommon::ImageViewInOut, MAX_TEXTURES + MAX_IMAGES> views; - std::array<GLuint, MAX_TEXTURES> samplers; + std::array<VideoCommon::SamplerId, MAX_TEXTURES> samplers; size_t views_index{}; - GLsizei sampler_binding{}; + size_t samplers_index{}; texture_cache.SynchronizeGraphicsDescriptors(); @@ -337,7 +337,6 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { for (u32 index = 0; index < desc.count; ++index) { const auto handle{read_handle(desc, index)}; views[views_index++] = {handle.first}; - samplers[sampler_binding++] = 0; } } } @@ -351,8 +350,8 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { const auto handle{read_handle(desc, index)}; views[views_index++] = {handle.first}; - Sampler* const sampler{texture_cache.GetGraphicsSampler(handle.second)}; - samplers[sampler_binding++] = sampler->Handle(); + VideoCommon::SamplerId sampler{texture_cache.GetGraphicsSamplerId(handle.second)}; + samplers[samplers_index++] = sampler; } } if constexpr (Spec::has_images) { @@ -445,10 +444,13 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { program_manager.BindSourcePrograms(source_programs); } const VideoCommon::ImageViewInOut* views_it{views.data()}; + const VideoCommon::SamplerId* samplers_it{samplers.data()}; GLsizei texture_binding = 0; GLsizei image_binding = 0; + GLsizei sampler_binding{}; std::array<GLuint, MAX_TEXTURES> textures; std::array<GLuint, MAX_IMAGES> images; + std::array<GLuint, MAX_TEXTURES> gl_samplers; const auto prepare_stage{[&](size_t stage) { buffer_cache.runtime.SetImagePointers(&textures[texture_binding], &images[image_binding]); buffer_cache.BindHostStageBuffers(stage); @@ -465,6 +467,13 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { u32 stage_image_binding{}; const auto& info{stage_infos[stage]}; + if constexpr (Spec::has_texture_buffers) { + for (const auto& desc : info.texture_buffer_descriptors) { + for (u32 index = 0; index < desc.count; ++index) { + gl_samplers[sampler_binding++] = 0; + } + } + } for (const auto& desc : info.texture_descriptors) { for (u32 index = 0; index < desc.count; ++index) { ImageView& image_view{texture_cache.GetImageView((views_it++)->id)}; @@ -474,6 +483,12 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { } ++texture_binding; ++stage_texture_binding; + + const Sampler& sampler{texture_cache.GetSampler(*(samplers_it++))}; + const bool use_fallback_sampler{sampler.HasAddedAnisotropy() && + !image_view.SupportsAnisotropy()}; + gl_samplers[sampler_binding++] = + use_fallback_sampler ? sampler.HandleWithDefaultAnisotropy() : sampler.Handle(); } } for (const auto& desc : info.image_descriptors) { @@ -534,7 +549,7 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { if (texture_binding != 0) { ASSERT(texture_binding == sampler_binding); glBindTextures(0, texture_binding, textures.data()); - glBindSamplers(0, sampler_binding, samplers.data()); + glBindSamplers(0, sampler_binding, gl_samplers.data()); } if (image_binding != 0) { glBindImageTextures(0, image_binding, images.data()); diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 6ecda2984..3f077311e 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -232,12 +232,14 @@ ShaderCache::ShaderCache(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindo .gl_max_compute_smem_size = device.GetMaxComputeSharedMemorySize(), }, host_info{ + .support_float64 = true, .support_float16 = false, .support_int64 = device.HasShaderInt64(), .needs_demote_reorder = device.IsAmd(), .support_snorm_render_buffer = false, .support_viewport_index_layer = device.HasVertexViewportLayer(), .support_geometry_shader_passthrough = device.HasGeometryShaderPassthrough(), + .support_conditional_barrier = device.SupportsConditionalBarriers(), } { if (use_asynchronous_shaders) { workers = CreateWorkers(); diff --git a/src/video_core/renderer_opengl/gl_shader_context.h b/src/video_core/renderer_opengl/gl_shader_context.h index 207a75d42..d12cd06fa 100644 --- a/src/video_core/renderer_opengl/gl_shader_context.h +++ b/src/video_core/renderer_opengl/gl_shader_context.h @@ -16,9 +16,9 @@ struct ShaderPools { inst.ReleaseContents(); } - Shader::ObjectPool<Shader::IR::Inst> inst; - Shader::ObjectPool<Shader::IR::Block> block; - Shader::ObjectPool<Shader::Maxwell::Flow::Block> flow_block; + Shader::ObjectPool<Shader::IR::Inst> inst{8192}; + Shader::ObjectPool<Shader::IR::Block> block{32}; + Shader::ObjectPool<Shader::Maxwell::Flow::Block> flow_block{32}; }; struct Context { diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index 1c5dbcdd8..3b446be07 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -1268,36 +1268,48 @@ Sampler::Sampler(TextureCacheRuntime& runtime, const TSCEntry& config) { UNIMPLEMENTED_IF(config.cubemap_anisotropy != 1); - sampler.Create(); - const GLuint handle = sampler.handle; - glSamplerParameteri(handle, GL_TEXTURE_WRAP_S, MaxwellToGL::WrapMode(config.wrap_u)); - glSamplerParameteri(handle, GL_TEXTURE_WRAP_T, MaxwellToGL::WrapMode(config.wrap_v)); - glSamplerParameteri(handle, GL_TEXTURE_WRAP_R, MaxwellToGL::WrapMode(config.wrap_p)); - glSamplerParameteri(handle, GL_TEXTURE_COMPARE_MODE, compare_mode); - glSamplerParameteri(handle, GL_TEXTURE_COMPARE_FUNC, compare_func); - glSamplerParameteri(handle, GL_TEXTURE_MAG_FILTER, mag); - glSamplerParameteri(handle, GL_TEXTURE_MIN_FILTER, min); - glSamplerParameterf(handle, GL_TEXTURE_LOD_BIAS, config.LodBias()); - glSamplerParameterf(handle, GL_TEXTURE_MIN_LOD, config.MinLod()); - glSamplerParameterf(handle, GL_TEXTURE_MAX_LOD, config.MaxLod()); - glSamplerParameterfv(handle, GL_TEXTURE_BORDER_COLOR, config.BorderColor().data()); - - if (GLAD_GL_ARB_texture_filter_anisotropic || GLAD_GL_EXT_texture_filter_anisotropic) { - const f32 max_anisotropy = std::clamp(config.MaxAnisotropy(), 1.0f, 16.0f); - glSamplerParameterf(handle, GL_TEXTURE_MAX_ANISOTROPY, max_anisotropy); - } else { - LOG_WARNING(Render_OpenGL, "GL_ARB_texture_filter_anisotropic is required"); - } - if (GLAD_GL_ARB_texture_filter_minmax || GLAD_GL_EXT_texture_filter_minmax) { - glSamplerParameteri(handle, GL_TEXTURE_REDUCTION_MODE_ARB, reduction_filter); - } else if (reduction_filter != GL_WEIGHTED_AVERAGE_ARB) { - LOG_WARNING(Render_OpenGL, "GL_ARB_texture_filter_minmax is required"); - } - if (GLAD_GL_ARB_seamless_cubemap_per_texture || GLAD_GL_AMD_seamless_cubemap_per_texture) { - glSamplerParameteri(handle, GL_TEXTURE_CUBE_MAP_SEAMLESS, seamless); - } else if (seamless == GL_FALSE) { - // We default to false because it's more common - LOG_WARNING(Render_OpenGL, "GL_ARB_seamless_cubemap_per_texture is required"); + const f32 max_anisotropy = std::clamp(config.MaxAnisotropy(), 1.0f, 16.0f); + + const auto create_sampler = [&](const f32 anisotropy) { + OGLSampler new_sampler; + new_sampler.Create(); + const GLuint handle = new_sampler.handle; + glSamplerParameteri(handle, GL_TEXTURE_WRAP_S, MaxwellToGL::WrapMode(config.wrap_u)); + glSamplerParameteri(handle, GL_TEXTURE_WRAP_T, MaxwellToGL::WrapMode(config.wrap_v)); + glSamplerParameteri(handle, GL_TEXTURE_WRAP_R, MaxwellToGL::WrapMode(config.wrap_p)); + glSamplerParameteri(handle, GL_TEXTURE_COMPARE_MODE, compare_mode); + glSamplerParameteri(handle, GL_TEXTURE_COMPARE_FUNC, compare_func); + glSamplerParameteri(handle, GL_TEXTURE_MAG_FILTER, mag); + glSamplerParameteri(handle, GL_TEXTURE_MIN_FILTER, min); + glSamplerParameterf(handle, GL_TEXTURE_LOD_BIAS, config.LodBias()); + glSamplerParameterf(handle, GL_TEXTURE_MIN_LOD, config.MinLod()); + glSamplerParameterf(handle, GL_TEXTURE_MAX_LOD, config.MaxLod()); + glSamplerParameterfv(handle, GL_TEXTURE_BORDER_COLOR, config.BorderColor().data()); + + if (GLAD_GL_ARB_texture_filter_anisotropic || GLAD_GL_EXT_texture_filter_anisotropic) { + glSamplerParameterf(handle, GL_TEXTURE_MAX_ANISOTROPY, anisotropy); + } else { + LOG_WARNING(Render_OpenGL, "GL_ARB_texture_filter_anisotropic is required"); + } + if (GLAD_GL_ARB_texture_filter_minmax || GLAD_GL_EXT_texture_filter_minmax) { + glSamplerParameteri(handle, GL_TEXTURE_REDUCTION_MODE_ARB, reduction_filter); + } else if (reduction_filter != GL_WEIGHTED_AVERAGE_ARB) { + LOG_WARNING(Render_OpenGL, "GL_ARB_texture_filter_minmax is required"); + } + if (GLAD_GL_ARB_seamless_cubemap_per_texture || GLAD_GL_AMD_seamless_cubemap_per_texture) { + glSamplerParameteri(handle, GL_TEXTURE_CUBE_MAP_SEAMLESS, seamless); + } else if (seamless == GL_FALSE) { + // We default to false because it's more common + LOG_WARNING(Render_OpenGL, "GL_ARB_seamless_cubemap_per_texture is required"); + } + return new_sampler; + }; + + sampler = create_sampler(max_anisotropy); + + const f32 max_anisotropy_default = static_cast<f32>(1U << config.max_anisotropy); + if (max_anisotropy > max_anisotropy_default) { + sampler_default_anisotropy = create_sampler(max_anisotropy_default); } } diff --git a/src/video_core/renderer_opengl/gl_texture_cache.h b/src/video_core/renderer_opengl/gl_texture_cache.h index 1148b73d7..3676eaaa9 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.h +++ b/src/video_core/renderer_opengl/gl_texture_cache.h @@ -309,12 +309,21 @@ class Sampler { public: explicit Sampler(TextureCacheRuntime&, const Tegra::Texture::TSCEntry&); - GLuint Handle() const noexcept { + [[nodiscard]] GLuint Handle() const noexcept { return sampler.handle; } + [[nodiscard]] GLuint HandleWithDefaultAnisotropy() const noexcept { + return sampler_default_anisotropy.handle; + } + + [[nodiscard]] bool HasAddedAnisotropy() const noexcept { + return static_cast<bool>(sampler_default_anisotropy.handle); + } + private: OGLSampler sampler; + OGLSampler sampler_default_anisotropy; }; class Framebuffer { diff --git a/src/video_core/renderer_vulkan/pipeline_helper.h b/src/video_core/renderer_vulkan/pipeline_helper.h index 983e1c2e1..71c783709 100644 --- a/src/video_core/renderer_vulkan/pipeline_helper.h +++ b/src/video_core/renderer_vulkan/pipeline_helper.h @@ -178,7 +178,7 @@ public: inline void PushImageDescriptors(TextureCache& texture_cache, GuestDescriptorQueue& guest_descriptor_queue, const Shader::Info& info, RescalingPushConstant& rescaling, - const VkSampler*& samplers, + const VideoCommon::SamplerId*& samplers, const VideoCommon::ImageViewInOut*& views) { const u32 num_texture_buffers = Shader::NumDescriptors(info.texture_buffer_descriptors); const u32 num_image_buffers = Shader::NumDescriptors(info.image_buffer_descriptors); @@ -187,10 +187,15 @@ inline void PushImageDescriptors(TextureCache& texture_cache, for (const auto& desc : info.texture_descriptors) { for (u32 index = 0; index < desc.count; ++index) { const VideoCommon::ImageViewId image_view_id{(views++)->id}; - const VkSampler sampler{*(samplers++)}; + const VideoCommon::SamplerId sampler_id{*(samplers++)}; ImageView& image_view{texture_cache.GetImageView(image_view_id)}; const VkImageView vk_image_view{image_view.Handle(desc.type)}; - guest_descriptor_queue.AddSampledImage(vk_image_view, sampler); + const Sampler& sampler{texture_cache.GetSampler(sampler_id)}; + const bool use_fallback_sampler{sampler.HasAddedAnisotropy() && + !image_view.SupportsAnisotropy()}; + const VkSampler vk_sampler{use_fallback_sampler ? sampler.HandleWithDefaultAnisotropy() + : sampler.Handle()}; + guest_descriptor_queue.AddSampledImage(vk_image_view, vk_sampler); rescaling.PushTexture(texture_cache.IsRescaling(image_view)); } } diff --git a/src/video_core/renderer_vulkan/vk_blit_screen.cpp b/src/video_core/renderer_vulkan/vk_blit_screen.cpp index 7cdde992b..acb143fc7 100644 --- a/src/video_core/renderer_vulkan/vk_blit_screen.cpp +++ b/src/video_core/renderer_vulkan/vk_blit_screen.cpp @@ -37,10 +37,6 @@ #include "video_core/vulkan_common/vulkan_memory_allocator.h" #include "video_core/vulkan_common/vulkan_wrapper.h" -#ifdef ANDROID -extern u32 GetAndroidScreenRotation(); -#endif - namespace Vulkan { namespace { @@ -78,47 +74,6 @@ struct ScreenRectVertex { } }; -#ifdef ANDROID - -std::array<f32, 4 * 4> MakeOrthographicMatrix(f32 width, f32 height) { - constexpr u32 ROTATION_0 = 0; - constexpr u32 ROTATION_90 = 1; - constexpr u32 ROTATION_180 = 2; - constexpr u32 ROTATION_270 = 3; - - // clang-format off - switch (GetAndroidScreenRotation()) { - case ROTATION_0: - // Desktop - return { 2.f / width, 0.f, 0.f, 0.f, - 0.f, 2.f / height, 0.f, 0.f, - 0.f, 0.f, 1.f, 0.f, - -1.f, -1.f, 0.f, 1.f}; - case ROTATION_180: - // Reverse desktop - return {-2.f / width, 0.f, 0.f, 0.f, - 0.f, -2.f / height, 0.f, 0.f, - 0.f, 0.f, 1.f, 0.f, - 1.f, 1.f, 0.f, 1.f}; - case ROTATION_270: - // Reverse landscape - return { 0.f, -2.f / width, 0.f, 0.f, - 2.f / height, 0.f, 0.f, 0.f, - 0.f, 0.f, 1.f, 0.f, - -1.f, 1.f, 0.f, 1.f}; - case ROTATION_90: - default: - // Landscape - return { 0.f, 2.f / width, 0.f, 0.f, - -2.f / height, 0.f, 0.f, 0.f, - 0.f, 0.f, 1.f, 0.f, - 1.f, -1.f, 0.f, 1.f}; - } - // clang-format on -} - -#else - std::array<f32, 4 * 4> MakeOrthographicMatrix(f32 width, f32 height) { // clang-format off return { 2.f / width, 0.f, 0.f, 0.f, @@ -128,8 +83,6 @@ std::array<f32, 4 * 4> MakeOrthographicMatrix(f32 width, f32 height) { // clang-format on } -#endif - u32 GetBytesPerPixel(const Tegra::FramebufferConfig& framebuffer) { using namespace VideoCore::Surface; return BytesPerBlock(PixelFormatFromGPUPixelFormat(framebuffer.pixel_format)); @@ -1159,7 +1112,7 @@ void BlitScreen::CreateRawImages(const Tegra::FramebufferConfig& framebuffer) { .pNext = nullptr, .flags = 0, .imageType = VK_IMAGE_TYPE_2D, - .format = GetFormat(framebuffer), + .format = used_on_framebuffer ? VK_FORMAT_R16G16B16A16_SFLOAT : GetFormat(framebuffer), .extent = { .width = (up_scale * framebuffer.width) >> down_shift, @@ -1180,14 +1133,14 @@ void BlitScreen::CreateRawImages(const Tegra::FramebufferConfig& framebuffer) { const auto create_commit = [&](vk::Image& image) { return memory_allocator.Commit(image, MemoryUsage::DeviceLocal); }; - const auto create_image_view = [&](vk::Image& image) { + const auto create_image_view = [&](vk::Image& image, bool used_on_framebuffer = false) { return device.GetLogical().CreateImageView(VkImageViewCreateInfo{ .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, .pNext = nullptr, .flags = 0, .image = *image, .viewType = VK_IMAGE_VIEW_TYPE_2D, - .format = GetFormat(framebuffer), + .format = used_on_framebuffer ? VK_FORMAT_R16G16B16A16_SFLOAT : GetFormat(framebuffer), .components = { .r = VK_COMPONENT_SWIZZLE_IDENTITY, @@ -1217,7 +1170,7 @@ void BlitScreen::CreateRawImages(const Tegra::FramebufferConfig& framebuffer) { const u32 down_shift = Settings::values.resolution_info.down_shift; aa_image = create_image(true, up_scale, down_shift); aa_commit = create_commit(aa_image); - aa_image_view = create_image_view(aa_image); + aa_image_view = create_image_view(aa_image, true); VkExtent2D size{ .width = (up_scale * framebuffer.width) >> down_shift, .height = (up_scale * framebuffer.height) >> down_shift, diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp index daa128399..e30fcb1ed 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp @@ -7,7 +7,6 @@ #include <span> #include <vector> -#include "video_core/buffer_cache/buffer_cache.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" #include "video_core/renderer_vulkan/vk_buffer_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" @@ -502,6 +501,36 @@ void BufferCacheRuntime::BindVertexBuffer(u32 index, VkBuffer buffer, u32 offset } } +void BufferCacheRuntime::BindVertexBuffers(VideoCommon::HostBindings<Buffer>& bindings) { + boost::container::small_vector<VkBuffer, 32> buffer_handles; + for (u32 index = 0; index < bindings.buffers.size(); ++index) { + auto handle = bindings.buffers[index]->Handle(); + if (handle == VK_NULL_HANDLE) { + bindings.offsets[index] = 0; + bindings.sizes[index] = VK_WHOLE_SIZE; + if (!device.HasNullDescriptor()) { + ReserveNullBuffer(); + handle = *null_buffer; + } + } + buffer_handles.push_back(handle); + } + if (device.IsExtExtendedDynamicStateSupported()) { + scheduler.Record([bindings = std::move(bindings), + buffer_handles = std::move(buffer_handles)](vk::CommandBuffer cmdbuf) { + cmdbuf.BindVertexBuffers2EXT( + bindings.min_index, bindings.max_index - bindings.min_index, buffer_handles.data(), + bindings.offsets.data(), bindings.sizes.data(), bindings.strides.data()); + }); + } else { + scheduler.Record([bindings = std::move(bindings), + buffer_handles = std::move(buffer_handles)](vk::CommandBuffer cmdbuf) { + cmdbuf.BindVertexBuffers(bindings.min_index, bindings.max_index - bindings.min_index, + buffer_handles.data(), bindings.offsets.data()); + }); + } +} + void BufferCacheRuntime::BindTransformFeedbackBuffer(u32 index, VkBuffer buffer, u32 offset, u32 size) { if (!device.IsExtTransformFeedbackSupported()) { @@ -523,6 +552,23 @@ void BufferCacheRuntime::BindTransformFeedbackBuffer(u32 index, VkBuffer buffer, }); } +void BufferCacheRuntime::BindTransformFeedbackBuffers(VideoCommon::HostBindings<Buffer>& bindings) { + if (!device.IsExtTransformFeedbackSupported()) { + // Already logged in the rasterizer + return; + } + boost::container::small_vector<VkBuffer, 4> buffer_handles; + for (u32 index = 0; index < bindings.buffers.size(); ++index) { + buffer_handles.push_back(bindings.buffers[index]->Handle()); + } + scheduler.Record([bindings = std::move(bindings), + buffer_handles = std::move(buffer_handles)](vk::CommandBuffer cmdbuf) { + cmdbuf.BindTransformFeedbackBuffersEXT(0, static_cast<u32>(buffer_handles.size()), + buffer_handles.data(), bindings.offsets.data(), + bindings.sizes.data()); + }); +} + void BufferCacheRuntime::ReserveNullBuffer() { if (null_buffer) { return; diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h index 92b4f7859..cdeef8846 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.h +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h @@ -18,6 +18,7 @@ namespace Vulkan { class Device; class DescriptorPool; class Scheduler; +struct HostVertexBinding; class BufferCacheRuntime; @@ -97,8 +98,12 @@ public: void BindVertexBuffer(u32 index, VkBuffer buffer, u32 offset, u32 size, u32 stride); + void BindVertexBuffers(VideoCommon::HostBindings<Buffer>& bindings); + void BindTransformFeedbackBuffer(u32 index, VkBuffer buffer, u32 offset, u32 size); + void BindTransformFeedbackBuffers(VideoCommon::HostBindings<Buffer>& bindings); + std::span<u8> BindMappedUniformBuffer([[maybe_unused]] size_t stage, [[maybe_unused]] u32 binding_index, u32 size) { const StagingBufferRef ref = staging_pool.Request(size, MemoryUsage::Upload); diff --git a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp index 733e70d9d..73e585c2b 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp @@ -115,7 +115,7 @@ void ComputePipeline::Configure(Tegra::Engines::KeplerCompute& kepler_compute, static constexpr size_t max_elements = 64; boost::container::static_vector<VideoCommon::ImageViewInOut, max_elements> views; - boost::container::static_vector<VkSampler, max_elements> samplers; + boost::container::static_vector<VideoCommon::SamplerId, max_elements> samplers; const auto& qmd{kepler_compute.launch_description}; const auto& cbufs{qmd.const_buffer_config}; @@ -160,8 +160,8 @@ void ComputePipeline::Configure(Tegra::Engines::KeplerCompute& kepler_compute, const auto handle{read_handle(desc, index)}; views.push_back({handle.first}); - Sampler* const sampler = texture_cache.GetComputeSampler(handle.second); - samplers.push_back(sampler->Handle()); + VideoCommon::SamplerId sampler = texture_cache.GetComputeSamplerId(handle.second); + samplers.push_back(sampler); } } for (const auto& desc : info.image_descriptors) { @@ -192,7 +192,7 @@ void ComputePipeline::Configure(Tegra::Engines::KeplerCompute& kepler_compute, buffer_cache.BindHostComputeBuffers(); RescalingPushConstant rescaling; - const VkSampler* samplers_it{samplers.data()}; + const VideoCommon::SamplerId* samplers_it{samplers.data()}; const VideoCommon::ImageViewInOut* views_it{views.data()}; PushImageDescriptors(texture_cache, guest_descriptor_queue, info, rescaling, samplers_it, views_it); diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp index 506b78f08..c1595642e 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp @@ -298,7 +298,7 @@ void GraphicsPipeline::AddTransition(GraphicsPipeline* transition) { template <typename Spec> void GraphicsPipeline::ConfigureImpl(bool is_indexed) { std::array<VideoCommon::ImageViewInOut, MAX_IMAGE_ELEMENTS> views; - std::array<VkSampler, MAX_IMAGE_ELEMENTS> samplers; + std::array<VideoCommon::SamplerId, MAX_IMAGE_ELEMENTS> samplers; size_t sampler_index{}; size_t view_index{}; @@ -367,8 +367,8 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { const auto handle{read_handle(desc, index)}; views[view_index++] = {handle.first}; - Sampler* const sampler{texture_cache.GetGraphicsSampler(handle.second)}; - samplers[sampler_index++] = sampler->Handle(); + VideoCommon::SamplerId sampler{texture_cache.GetGraphicsSamplerId(handle.second)}; + samplers[sampler_index++] = sampler; } } if constexpr (Spec::has_images) { @@ -453,7 +453,7 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { RescalingPushConstant rescaling; RenderAreaPushConstant render_area; - const VkSampler* samplers_it{samplers.data()}; + const VideoCommon::SamplerId* samplers_it{samplers.data()}; const VideoCommon::ImageViewInOut* views_it{views.data()}; const auto prepare_stage{[&](size_t stage) LAMBDA_FORCEINLINE { buffer_cache.BindHostStageBuffers(stage); diff --git a/src/video_core/renderer_vulkan/vk_master_semaphore.cpp b/src/video_core/renderer_vulkan/vk_master_semaphore.cpp index b128c4f6e..5eeda08d2 100644 --- a/src/video_core/renderer_vulkan/vk_master_semaphore.cpp +++ b/src/video_core/renderer_vulkan/vk_master_semaphore.cpp @@ -3,6 +3,7 @@ #include <thread> +#include "common/polyfill_ranges.h" #include "common/settings.h" #include "video_core/renderer_vulkan/vk_master_semaphore.h" #include "video_core/vulkan_common/vulkan_device.h" diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp index 9482e91b0..a2cfb2105 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp @@ -350,6 +350,7 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device .has_broken_spirv_subgroup_mask_vector_extract_dynamic = driver_id == VK_DRIVER_ID_QUALCOMM_PROPRIETARY}; host_info = Shader::HostTranslateInfo{ + .support_float64 = device.IsFloat64Supported(), .support_float16 = device.IsFloat16Supported(), .support_int64 = device.IsShaderInt64Supported(), .needs_demote_reorder = @@ -357,6 +358,7 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device .support_snorm_render_buffer = true, .support_viewport_index_layer = device.IsExtShaderViewportIndexLayerSupported(), .support_geometry_shader_passthrough = device.IsNvGeometryShaderPassthroughSupported(), + .support_conditional_barrier = device.SupportsConditionalBarriers(), }; if (device.GetMaxVertexInputAttributes() < Maxwell::NumVertexAttributes) { @@ -703,10 +705,7 @@ std::unique_ptr<ComputePipeline> PipelineCache::CreateComputePipeline( std::unique_ptr<ComputePipeline> PipelineCache::CreateComputePipeline( ShaderPools& pools, const ComputePipelineCacheKey& key, Shader::Environment& env, PipelineStatistics* statistics, bool build_in_parallel) try { - // TODO: Remove this when Intel fixes their shader compiler. - // https://github.com/IGCIT/Intel-GPU-Community-Issue-Tracker-IGCIT/issues/159 - if (device.GetDriverID() == VK_DRIVER_ID_INTEL_PROPRIETARY_WINDOWS && - !Settings::values.enable_compute_pipelines.GetValue()) { + if (device.HasBrokenCompute()) { LOG_ERROR(Render_Vulkan, "Skipping 0x{:016x}", key.Hash()); return nullptr; } diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.h b/src/video_core/renderer_vulkan/vk_pipeline_cache.h index 15aa7e224..e323ea0fd 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.h +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.h @@ -92,9 +92,9 @@ struct ShaderPools { inst.ReleaseContents(); } - Shader::ObjectPool<Shader::IR::Inst> inst; - Shader::ObjectPool<Shader::IR::Block> block; - Shader::ObjectPool<Shader::Maxwell::Flow::Block> flow_block; + Shader::ObjectPool<Shader::IR::Inst> inst{8192}; + Shader::ObjectPool<Shader::IR::Block> block{32}; + Shader::ObjectPool<Shader::Maxwell::Flow::Block> flow_block{32}; }; class PipelineCache : public VideoCommon::ShaderCache { diff --git a/src/video_core/renderer_vulkan/vk_swapchain.cpp b/src/video_core/renderer_vulkan/vk_swapchain.cpp index afcf34fba..d3cddac69 100644 --- a/src/video_core/renderer_vulkan/vk_swapchain.cpp +++ b/src/video_core/renderer_vulkan/vk_swapchain.cpp @@ -231,7 +231,12 @@ void Swapchain::CreateSwapchain(const VkSurfaceCapabilitiesKHR& capabilities, bo .imageSharingMode = VK_SHARING_MODE_EXCLUSIVE, .queueFamilyIndexCount = 0, .pQueueFamilyIndices = nullptr, +#ifdef ANDROID + // On Android, do not allow surface rotation to deviate from the frontend. + .preTransform = VK_SURFACE_TRANSFORM_IDENTITY_BIT_KHR, +#else .preTransform = capabilities.currentTransform, +#endif .compositeAlpha = alpha_flags, .presentMode = present_mode, .clipped = VK_FALSE, diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index 8711e2a87..f025f618b 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -1802,27 +1802,36 @@ Sampler::Sampler(TextureCacheRuntime& runtime, const Tegra::Texture::TSCEntry& t // Some games have samplers with garbage. Sanitize them here. const f32 max_anisotropy = std::clamp(tsc.MaxAnisotropy(), 1.0f, 16.0f); - sampler = device.GetLogical().CreateSampler(VkSamplerCreateInfo{ - .sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO, - .pNext = pnext, - .flags = 0, - .magFilter = MaxwellToVK::Sampler::Filter(tsc.mag_filter), - .minFilter = MaxwellToVK::Sampler::Filter(tsc.min_filter), - .mipmapMode = MaxwellToVK::Sampler::MipmapMode(tsc.mipmap_filter), - .addressModeU = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_u, tsc.mag_filter), - .addressModeV = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_v, tsc.mag_filter), - .addressModeW = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_p, tsc.mag_filter), - .mipLodBias = tsc.LodBias(), - .anisotropyEnable = static_cast<VkBool32>(max_anisotropy > 1.0f ? VK_TRUE : VK_FALSE), - .maxAnisotropy = max_anisotropy, - .compareEnable = tsc.depth_compare_enabled, - .compareOp = MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func), - .minLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.0f : tsc.MinLod(), - .maxLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.25f : tsc.MaxLod(), - .borderColor = - arbitrary_borders ? VK_BORDER_COLOR_FLOAT_CUSTOM_EXT : ConvertBorderColor(color), - .unnormalizedCoordinates = VK_FALSE, - }); + const auto create_sampler = [&](const f32 anisotropy) { + return device.GetLogical().CreateSampler(VkSamplerCreateInfo{ + .sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO, + .pNext = pnext, + .flags = 0, + .magFilter = MaxwellToVK::Sampler::Filter(tsc.mag_filter), + .minFilter = MaxwellToVK::Sampler::Filter(tsc.min_filter), + .mipmapMode = MaxwellToVK::Sampler::MipmapMode(tsc.mipmap_filter), + .addressModeU = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_u, tsc.mag_filter), + .addressModeV = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_v, tsc.mag_filter), + .addressModeW = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_p, tsc.mag_filter), + .mipLodBias = tsc.LodBias(), + .anisotropyEnable = static_cast<VkBool32>(anisotropy > 1.0f ? VK_TRUE : VK_FALSE), + .maxAnisotropy = anisotropy, + .compareEnable = tsc.depth_compare_enabled, + .compareOp = MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func), + .minLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.0f : tsc.MinLod(), + .maxLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.25f : tsc.MaxLod(), + .borderColor = + arbitrary_borders ? VK_BORDER_COLOR_FLOAT_CUSTOM_EXT : ConvertBorderColor(color), + .unnormalizedCoordinates = VK_FALSE, + }); + }; + + sampler = create_sampler(max_anisotropy); + + const f32 max_anisotropy_default = static_cast<f32>(1U << tsc.max_anisotropy); + if (max_anisotropy > max_anisotropy_default) { + sampler_default_anisotropy = create_sampler(max_anisotropy_default); + } } Framebuffer::Framebuffer(TextureCacheRuntime& runtime, std::span<ImageView*, NUM_RT> color_buffers, diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.h b/src/video_core/renderer_vulkan/vk_texture_cache.h index 0f7a5ffd4..f14525dcb 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.h +++ b/src/video_core/renderer_vulkan/vk_texture_cache.h @@ -279,8 +279,17 @@ public: return *sampler; } + [[nodiscard]] VkSampler HandleWithDefaultAnisotropy() const noexcept { + return *sampler_default_anisotropy; + } + + [[nodiscard]] bool HasAddedAnisotropy() const noexcept { + return static_cast<bool>(sampler_default_anisotropy); + } + private: vk::Sampler sampler; + vk::Sampler sampler_default_anisotropy; }; class Framebuffer { diff --git a/src/video_core/texture_cache/image_info.cpp b/src/video_core/texture_cache/image_info.cpp index e8ddde691..b72788c6d 100644 --- a/src/video_core/texture_cache/image_info.cpp +++ b/src/video_core/texture_cache/image_info.cpp @@ -22,6 +22,9 @@ using Tegra::Texture::TICEntry; using VideoCore::Surface::PixelFormat; using VideoCore::Surface::SurfaceType; +constexpr u32 RescaleHeightThreshold = 288; +constexpr u32 DownscaleHeightThreshold = 512; + ImageInfo::ImageInfo(const TICEntry& config) noexcept { forced_flushed = config.IsPitchLinear() && !Settings::values.use_reactive_flushing.GetValue(); dma_downloaded = forced_flushed; @@ -113,8 +116,9 @@ ImageInfo::ImageInfo(const TICEntry& config) noexcept { layer_stride = CalculateLayerStride(*this); maybe_unaligned_layer_stride = CalculateLayerSize(*this); rescaleable &= (block.depth == 0) && resources.levels == 1; - rescaleable &= size.height > 256 || GetFormatType(format) != SurfaceType::ColorTexture; - downscaleable = size.height > 512; + rescaleable &= size.height > RescaleHeightThreshold || + GetFormatType(format) != SurfaceType::ColorTexture; + downscaleable = size.height > DownscaleHeightThreshold; } } @@ -152,8 +156,8 @@ ImageInfo::ImageInfo(const Maxwell3D::Regs::RenderTargetConfig& ct, size.depth = ct.depth; } else { rescaleable = block.depth == 0; - rescaleable &= size.height > 256; - downscaleable = size.height > 512; + rescaleable &= size.height > RescaleHeightThreshold; + downscaleable = size.height > DownscaleHeightThreshold; type = ImageType::e2D; resources.layers = ct.depth; } @@ -232,8 +236,8 @@ ImageInfo::ImageInfo(const Fermi2D::Surface& config) noexcept { .height = config.height, .depth = 1, }; - rescaleable = block.depth == 0 && size.height > 256; - downscaleable = size.height > 512; + rescaleable = block.depth == 0 && size.height > RescaleHeightThreshold; + downscaleable = size.height > DownscaleHeightThreshold; } } @@ -275,8 +279,8 @@ ImageInfo::ImageInfo(const Tegra::DMA::ImageOperand& config) noexcept { resources.layers = 1; layer_stride = CalculateLayerStride(*this); maybe_unaligned_layer_stride = CalculateLayerSize(*this); - rescaleable = block.depth == 0 && size.height > 256; - downscaleable = size.height > 512; + rescaleable = block.depth == 0 && size.height > RescaleHeightThreshold; + downscaleable = size.height > DownscaleHeightThreshold; } } // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_view_base.cpp b/src/video_core/texture_cache/image_view_base.cpp index d134b6738..0c5f4450d 100644 --- a/src/video_core/texture_cache/image_view_base.cpp +++ b/src/video_core/texture_cache/image_view_base.cpp @@ -45,4 +45,56 @@ ImageViewBase::ImageViewBase(const ImageInfo& info, const ImageViewInfo& view_in ImageViewBase::ImageViewBase(const NullImageViewParams&) : image_id{NULL_IMAGE_ID} {} +bool ImageViewBase::SupportsAnisotropy() const noexcept { + const bool has_mips = range.extent.levels > 1; + const bool is_2d = type == ImageViewType::e2D || type == ImageViewType::e2DArray; + if (!has_mips || !is_2d) { + return false; + } + + switch (format) { + case PixelFormat::R8_UNORM: + case PixelFormat::R8_SNORM: + case PixelFormat::R8_SINT: + case PixelFormat::R8_UINT: + case PixelFormat::BC4_UNORM: + case PixelFormat::BC4_SNORM: + case PixelFormat::BC5_UNORM: + case PixelFormat::BC5_SNORM: + case PixelFormat::R32G32_FLOAT: + case PixelFormat::R32G32_SINT: + case PixelFormat::R32_FLOAT: + case PixelFormat::R16_FLOAT: + case PixelFormat::R16_UNORM: + case PixelFormat::R16_SNORM: + case PixelFormat::R16_UINT: + case PixelFormat::R16_SINT: + case PixelFormat::R16G16_UNORM: + case PixelFormat::R16G16_FLOAT: + case PixelFormat::R16G16_UINT: + case PixelFormat::R16G16_SINT: + case PixelFormat::R16G16_SNORM: + case PixelFormat::R8G8_UNORM: + case PixelFormat::R8G8_SNORM: + case PixelFormat::R8G8_SINT: + case PixelFormat::R8G8_UINT: + case PixelFormat::R32G32_UINT: + case PixelFormat::R32_UINT: + case PixelFormat::R32_SINT: + case PixelFormat::G4R4_UNORM: + // Depth formats + case PixelFormat::D32_FLOAT: + case PixelFormat::D16_UNORM: + // Stencil formats + case PixelFormat::S8_UINT: + // DepthStencil formats + case PixelFormat::D24_UNORM_S8_UINT: + case PixelFormat::S8_UINT_D24_UNORM: + case PixelFormat::D32_FLOAT_S8_UINT: + return false; + default: + return true; + } +} + } // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_view_base.h b/src/video_core/texture_cache/image_view_base.h index a25ae1d4a..87549ffff 100644 --- a/src/video_core/texture_cache/image_view_base.h +++ b/src/video_core/texture_cache/image_view_base.h @@ -33,6 +33,8 @@ struct ImageViewBase { return type == ImageViewType::Buffer; } + [[nodiscard]] bool SupportsAnisotropy() const noexcept; + ImageId image_id{}; GPUVAddr gpu_addr = 0; PixelFormat format{}; diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index c7f7448e9..d58bb69ff 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -186,6 +186,10 @@ void TextureCache<P>::FillComputeImageViews(std::span<ImageViewInOut> views) { template <class P> void TextureCache<P>::CheckFeedbackLoop(std::span<const ImageViewInOut> views) { + if (!Settings::values.barrier_feedback_loops.GetValue()) { + return; + } + const bool requires_barrier = [&] { for (const auto& view : views) { if (!view.id) { @@ -222,30 +226,50 @@ void TextureCache<P>::CheckFeedbackLoop(std::span<const ImageViewInOut> views) { template <class P> typename P::Sampler* TextureCache<P>::GetGraphicsSampler(u32 index) { + return &slot_samplers[GetGraphicsSamplerId(index)]; +} + +template <class P> +typename P::Sampler* TextureCache<P>::GetComputeSampler(u32 index) { + return &slot_samplers[GetComputeSamplerId(index)]; +} + +template <class P> +SamplerId TextureCache<P>::GetGraphicsSamplerId(u32 index) { if (index > channel_state->graphics_sampler_table.Limit()) { LOG_DEBUG(HW_GPU, "Invalid sampler index={}", index); - return &slot_samplers[NULL_SAMPLER_ID]; + return NULL_SAMPLER_ID; } const auto [descriptor, is_new] = channel_state->graphics_sampler_table.Read(index); SamplerId& id = channel_state->graphics_sampler_ids[index]; if (is_new) { id = FindSampler(descriptor); } - return &slot_samplers[id]; + return id; } template <class P> -typename P::Sampler* TextureCache<P>::GetComputeSampler(u32 index) { +SamplerId TextureCache<P>::GetComputeSamplerId(u32 index) { if (index > channel_state->compute_sampler_table.Limit()) { LOG_DEBUG(HW_GPU, "Invalid sampler index={}", index); - return &slot_samplers[NULL_SAMPLER_ID]; + return NULL_SAMPLER_ID; } const auto [descriptor, is_new] = channel_state->compute_sampler_table.Read(index); SamplerId& id = channel_state->compute_sampler_ids[index]; if (is_new) { id = FindSampler(descriptor); } - return &slot_samplers[id]; + return id; +} + +template <class P> +const typename P::Sampler& TextureCache<P>::GetSampler(SamplerId id) const noexcept { + return slot_samplers[id]; +} + +template <class P> +typename P::Sampler& TextureCache<P>::GetSampler(SamplerId id) noexcept { + return slot_samplers[id]; } template <class P> @@ -280,7 +304,7 @@ void TextureCache<P>::SynchronizeComputeDescriptors() { } template <class P> -bool TextureCache<P>::RescaleRenderTargets(bool is_clear) { +bool TextureCache<P>::RescaleRenderTargets() { auto& flags = maxwell3d->dirty.flags; u32 scale_rating = 0; bool rescaled = false; @@ -318,13 +342,13 @@ bool TextureCache<P>::RescaleRenderTargets(bool is_clear) { ImageViewId& color_buffer_id = render_targets.color_buffer_ids[index]; if (flags[Dirty::ColorBuffer0 + index] || force) { flags[Dirty::ColorBuffer0 + index] = false; - BindRenderTarget(&color_buffer_id, FindColorBuffer(index, is_clear)); + BindRenderTarget(&color_buffer_id, FindColorBuffer(index)); } check_rescale(color_buffer_id, tmp_color_images[index]); } if (flags[Dirty::ZetaBuffer] || force) { flags[Dirty::ZetaBuffer] = false; - BindRenderTarget(&render_targets.depth_buffer_id, FindDepthBuffer(is_clear)); + BindRenderTarget(&render_targets.depth_buffer_id, FindDepthBuffer()); } check_rescale(render_targets.depth_buffer_id, tmp_depth_image); @@ -389,7 +413,7 @@ void TextureCache<P>::UpdateRenderTargets(bool is_clear) { return; } - const bool rescaled = RescaleRenderTargets(is_clear); + const bool rescaled = RescaleRenderTargets(); if (is_rescaling != rescaled) { flags[Dirty::RescaleViewports] = true; flags[Dirty::RescaleScissors] = true; @@ -1658,7 +1682,7 @@ SamplerId TextureCache<P>::FindSampler(const TSCEntry& config) { } template <class P> -ImageViewId TextureCache<P>::FindColorBuffer(size_t index, bool is_clear) { +ImageViewId TextureCache<P>::FindColorBuffer(size_t index) { const auto& regs = maxwell3d->regs; if (index >= regs.rt_control.count) { return ImageViewId{}; @@ -1672,11 +1696,11 @@ ImageViewId TextureCache<P>::FindColorBuffer(size_t index, bool is_clear) { return ImageViewId{}; } const ImageInfo info(regs.rt[index], regs.anti_alias_samples_mode); - return FindRenderTargetView(info, gpu_addr, is_clear); + return FindRenderTargetView(info, gpu_addr); } template <class P> -ImageViewId TextureCache<P>::FindDepthBuffer(bool is_clear) { +ImageViewId TextureCache<P>::FindDepthBuffer() { const auto& regs = maxwell3d->regs; if (!regs.zeta_enable) { return ImageViewId{}; @@ -1686,18 +1710,16 @@ ImageViewId TextureCache<P>::FindDepthBuffer(bool is_clear) { return ImageViewId{}; } const ImageInfo info(regs.zeta, regs.zeta_size, regs.anti_alias_samples_mode); - return FindRenderTargetView(info, gpu_addr, is_clear); + return FindRenderTargetView(info, gpu_addr); } template <class P> -ImageViewId TextureCache<P>::FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr, - bool is_clear) { - const auto options = is_clear ? RelaxedOptions::Samples : RelaxedOptions{}; +ImageViewId TextureCache<P>::FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr) { ImageId image_id{}; bool delete_state = has_deleted_images; do { has_deleted_images = false; - image_id = FindOrInsertImage(info, gpu_addr, options); + image_id = FindOrInsertImage(info, gpu_addr); delete_state |= has_deleted_images; } while (has_deleted_images); has_deleted_images = delete_state; diff --git a/src/video_core/texture_cache/texture_cache_base.h b/src/video_core/texture_cache/texture_cache_base.h index 3bfa92154..44232b961 100644 --- a/src/video_core/texture_cache/texture_cache_base.h +++ b/src/video_core/texture_cache/texture_cache_base.h @@ -159,6 +159,18 @@ public: /// Get the sampler from the compute descriptor table in the specified index Sampler* GetComputeSampler(u32 index); + /// Get the sampler id from the graphics descriptor table in the specified index + SamplerId GetGraphicsSamplerId(u32 index); + + /// Get the sampler id from the compute descriptor table in the specified index + SamplerId GetComputeSamplerId(u32 index); + + /// Return a constant reference to the given sampler id + [[nodiscard]] const Sampler& GetSampler(SamplerId id) const noexcept; + + /// Return a reference to the given sampler id + [[nodiscard]] Sampler& GetSampler(SamplerId id) noexcept; + /// Refresh the state for graphics image view and sampler descriptors void SynchronizeGraphicsDescriptors(); @@ -166,9 +178,8 @@ public: void SynchronizeComputeDescriptors(); /// Updates the Render Targets if they can be rescaled - /// @param is_clear True when the render targets are being used for clears /// @retval True if the Render Targets have been rescaled. - bool RescaleRenderTargets(bool is_clear); + bool RescaleRenderTargets(); /// Update bound render targets and upload memory if necessary /// @param is_clear True when the render targets are being used for clears @@ -324,14 +335,13 @@ private: [[nodiscard]] SamplerId FindSampler(const TSCEntry& config); /// Find or create an image view for the given color buffer index - [[nodiscard]] ImageViewId FindColorBuffer(size_t index, bool is_clear); + [[nodiscard]] ImageViewId FindColorBuffer(size_t index); /// Find or create an image view for the depth buffer - [[nodiscard]] ImageViewId FindDepthBuffer(bool is_clear); + [[nodiscard]] ImageViewId FindDepthBuffer(); /// Find or create a view for a render target with the given image parameters - [[nodiscard]] ImageViewId FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr, - bool is_clear); + [[nodiscard]] ImageViewId FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr); /// Iterates over all the images in a region calling func template <typename Func> diff --git a/src/video_core/textures/texture.cpp b/src/video_core/textures/texture.cpp index 4a80a59f9..d8b88d9bc 100644 --- a/src/video_core/textures/texture.cpp +++ b/src/video_core/textures/texture.cpp @@ -62,7 +62,12 @@ std::array<float, 4> TSCEntry::BorderColor() const noexcept { } float TSCEntry::MaxAnisotropy() const noexcept { - if (max_anisotropy == 0 && mipmap_filter != TextureMipmapFilter::Linear) { + const bool is_suitable_mipmap_filter = mipmap_filter != TextureMipmapFilter::None; + const bool has_regular_lods = min_lod_clamp == 0 && max_lod_clamp >= 256; + const bool is_bilinear_filter = min_filter == TextureFilter::Linear && + reduction_filter == SamplerReduction::WeightedAverage; + if (max_anisotropy == 0 && (!is_suitable_mipmap_filter || !has_regular_lods || + !is_bilinear_filter || depth_compare_enabled)) { return 1.0f; } const auto anisotropic_settings = Settings::values.max_anisotropy.GetValue(); diff --git a/src/video_core/vulkan_common/vulkan_device.cpp b/src/video_core/vulkan_common/vulkan_device.cpp index 0158b6b0d..fa9cde75b 100644 --- a/src/video_core/vulkan_common/vulkan_device.cpp +++ b/src/video_core/vulkan_common/vulkan_device.cpp @@ -344,6 +344,7 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR const bool is_qualcomm = driver_id == VK_DRIVER_ID_QUALCOMM_PROPRIETARY; const bool is_turnip = driver_id == VK_DRIVER_ID_MESA_TURNIP; const bool is_s8gen2 = device_id == 0x43050a01; + const bool is_arm = driver_id == VK_DRIVER_ID_ARM_PROPRIETARY; if ((is_mvk || is_qualcomm || is_turnip) && !is_suitable) { LOG_WARNING(Render_Vulkan, "Unsuitable driver, continuing anyway"); @@ -386,10 +387,11 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR IsFormatSupported(VK_FORMAT_D24_UNORM_S8_UINT, VK_FORMAT_FEATURE_DEPTH_STENCIL_ATTACHMENT_BIT, FormatType::Optimal); + supports_conditional_barriers = !(is_intel_anv || is_intel_windows); + CollectPhysicalMemoryInfo(); CollectToolingInfo(); -#ifdef ANDROID if (is_qualcomm || is_turnip) { LOG_WARNING(Render_Vulkan, "Qualcomm and Turnip drivers have broken VK_EXT_custom_border_color"); @@ -409,7 +411,7 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR extensions.push_descriptor = false; loaded_extensions.erase(VK_KHR_PUSH_DESCRIPTOR_EXTENSION_NAME); -#ifdef ARCHITECTURE_arm64 +#if defined(ANDROID) && defined(ARCHITECTURE_arm64) // Patch the driver to enable BCn textures. const auto major = (properties.properties.driverVersion >> 24) << 2; const auto minor = (properties.properties.driverVersion >> 12) & 0xFFFU; @@ -429,18 +431,23 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR } else { LOG_WARNING(Render_Vulkan, "Adreno driver can't be patched to enable BCn textures"); } -#endif // ARCHITECTURE_arm64 +#endif } - const bool is_arm = driver_id == VK_DRIVER_ID_ARM_PROPRIETARY; if (is_arm) { must_emulate_scaled_formats = true; LOG_WARNING(Render_Vulkan, "ARM drivers have broken VK_EXT_extended_dynamic_state"); extensions.extended_dynamic_state = false; loaded_extensions.erase(VK_EXT_EXTENDED_DYNAMIC_STATE_EXTENSION_NAME); + + LOG_WARNING(Render_Vulkan, "ARM drivers have broken VK_EXT_extended_dynamic_state2"); + features.extended_dynamic_state2.extendedDynamicState2 = false; + features.extended_dynamic_state2.extendedDynamicState2LogicOp = false; + features.extended_dynamic_state2.extendedDynamicState2PatchControlPoints = false; + extensions.extended_dynamic_state2 = false; + loaded_extensions.erase(VK_EXT_EXTENDED_DYNAMIC_STATE_2_EXTENSION_NAME); } -#endif // ANDROID if (is_nvidia) { const u32 nv_major_version = (properties.properties.driverVersion >> 22) & 0x3ff; @@ -555,6 +562,9 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR LOG_WARNING(Render_Vulkan, "Intel proprietary drivers do not support MSAA image blits"); cant_blit_msaa = true; } + has_broken_compute = + CheckBrokenCompute(properties.driver.driverID, properties.properties.driverVersion) && + !Settings::values.enable_compute_pipelines.GetValue(); if (is_intel_anv || (is_qualcomm && !is_s8gen2)) { LOG_WARNING(Render_Vulkan, "Driver does not support native BGR format"); must_emulate_bgr565 = true; @@ -776,9 +786,6 @@ bool Device::GetSuitability(bool requires_swapchain) { FOR_EACH_VK_FEATURE_EXT(FEATURE_EXTENSION); FOR_EACH_VK_EXTENSION(EXTENSION); -#ifdef _WIN32 - FOR_EACH_VK_EXTENSION_WIN32(EXTENSION); -#endif #undef FEATURE_EXTENSION #undef EXTENSION @@ -797,11 +804,6 @@ bool Device::GetSuitability(bool requires_swapchain) { FOR_EACH_VK_RECOMMENDED_EXTENSION(LOG_EXTENSION); FOR_EACH_VK_MANDATORY_EXTENSION(CHECK_EXTENSION); -#ifdef _WIN32 - FOR_EACH_VK_MANDATORY_EXTENSION_WIN32(CHECK_EXTENSION); -#else - FOR_EACH_VK_MANDATORY_EXTENSION_GENERIC(CHECK_EXTENSION); -#endif if (requires_swapchain) { CHECK_EXTENSION(VK_KHR_SWAPCHAIN_EXTENSION_NAME); diff --git a/src/video_core/vulkan_common/vulkan_device.h b/src/video_core/vulkan_common/vulkan_device.h index d62a103a1..0b634a876 100644 --- a/src/video_core/vulkan_common/vulkan_device.h +++ b/src/video_core/vulkan_common/vulkan_device.h @@ -10,6 +10,7 @@ #include <vector> #include "common/common_types.h" +#include "common/logging/log.h" #include "common/settings.h" #include "video_core/vulkan_common/vulkan_wrapper.h" @@ -68,7 +69,6 @@ EXTENSION(EXT, VERTEX_ATTRIBUTE_DIVISOR, vertex_attribute_divisor) \ EXTENSION(KHR, DRAW_INDIRECT_COUNT, draw_indirect_count) \ EXTENSION(KHR, DRIVER_PROPERTIES, driver_properties) \ - EXTENSION(KHR, EXTERNAL_MEMORY_FD, external_memory_fd) \ EXTENSION(KHR, PUSH_DESCRIPTOR, push_descriptor) \ EXTENSION(KHR, SAMPLER_MIRROR_CLAMP_TO_EDGE, sampler_mirror_clamp_to_edge) \ EXTENSION(KHR, SHADER_FLOAT_CONTROLS, shader_float_controls) \ @@ -80,9 +80,6 @@ EXTENSION(NV, VIEWPORT_ARRAY2, viewport_array2) \ EXTENSION(NV, VIEWPORT_SWIZZLE, viewport_swizzle) -#define FOR_EACH_VK_EXTENSION_WIN32(EXTENSION) \ - EXTENSION(KHR, EXTERNAL_MEMORY_WIN32, external_memory_win32) - // Define extensions which must be supported. #define FOR_EACH_VK_MANDATORY_EXTENSION(EXTENSION_NAME) \ EXTENSION_NAME(VK_EXT_VERTEX_ATTRIBUTE_DIVISOR_EXTENSION_NAME) \ @@ -90,12 +87,6 @@ EXTENSION_NAME(VK_KHR_SAMPLER_MIRROR_CLAMP_TO_EDGE_EXTENSION_NAME) \ EXTENSION_NAME(VK_KHR_SHADER_FLOAT_CONTROLS_EXTENSION_NAME) -#define FOR_EACH_VK_MANDATORY_EXTENSION_GENERIC(EXTENSION_NAME) \ - EXTENSION_NAME(VK_KHR_EXTERNAL_MEMORY_FD_EXTENSION_NAME) - -#define FOR_EACH_VK_MANDATORY_EXTENSION_WIN32(EXTENSION_NAME) \ - EXTENSION_NAME(VK_KHR_EXTERNAL_MEMORY_WIN32_EXTENSION_NAME) - // Define extensions where the absence of the extension may result in a degraded experience. #define FOR_EACH_VK_RECOMMENDED_EXTENSION(EXTENSION_NAME) \ EXTENSION_NAME(VK_EXT_CONSERVATIVE_RASTERIZATION_EXTENSION_NAME) \ @@ -300,6 +291,11 @@ public: return GetDriverID() != VK_DRIVER_ID_QUALCOMM_PROPRIETARY; } + /// Returns true if the device suppors float64 natively. + bool IsFloat64Supported() const { + return features.features.shaderFloat64; + } + /// Returns true if the device supports float16 natively. bool IsFloat16Supported() const { return features.shader_float16_int8.shaderFloat16; @@ -523,6 +519,11 @@ public: return has_renderdoc || has_nsight_graphics || Settings::values.renderer_debug.GetValue(); } + /// @returns True if compute pipelines can cause crashing. + bool HasBrokenCompute() const { + return has_broken_compute; + } + /// Returns true when the device does not properly support cube compatibility. bool HasBrokenCubeImageCompability() const { return has_broken_cube_compatibility; @@ -580,6 +581,26 @@ public: return properties.properties.limits.maxVertexInputBindings; } + bool SupportsConditionalBarriers() const { + return supports_conditional_barriers; + } + + [[nodiscard]] static constexpr bool CheckBrokenCompute(VkDriverId driver_id, + u32 driver_version) { + if (driver_id == VK_DRIVER_ID_INTEL_PROPRIETARY_WINDOWS) { + const u32 major = VK_API_VERSION_MAJOR(driver_version); + const u32 minor = VK_API_VERSION_MINOR(driver_version); + const u32 patch = VK_API_VERSION_PATCH(driver_version); + if (major == 0 && minor == 405 && patch < 286) { + LOG_WARNING( + Render_Vulkan, + "Intel proprietary drivers 0.405.0 until 0.405.286 have broken compute"); + return true; + } + } + return false; + } + private: /// Checks if the physical device is suitable and configures the object state /// with all necessary info about its properties. @@ -627,7 +648,6 @@ private: FOR_EACH_VK_FEATURE_1_3(FEATURE); FOR_EACH_VK_FEATURE_EXT(FEATURE); FOR_EACH_VK_EXTENSION(EXTENSION); - FOR_EACH_VK_EXTENSION_WIN32(EXTENSION); #undef EXTENSION #undef FEATURE @@ -674,6 +694,7 @@ private: bool is_integrated{}; ///< Is GPU an iGPU. bool is_virtual{}; ///< Is GPU a virtual GPU. bool is_non_gpu{}; ///< Is SoftwareRasterizer, FPGA, non-GPU device. + bool has_broken_compute{}; ///< Compute shaders can cause crashes bool has_broken_cube_compatibility{}; ///< Has broken cube compatibility bit bool has_renderdoc{}; ///< Has RenderDoc attached bool has_nsight_graphics{}; ///< Has Nsight Graphics attached @@ -683,6 +704,7 @@ private: bool must_emulate_bgr565{}; ///< Emulates BGR565 by swizzling RGB565 format. bool dynamic_state3_blending{}; ///< Has all blending features of dynamic_state3. bool dynamic_state3_enables{}; ///< Has all enables features of dynamic_state3. + bool supports_conditional_barriers{}; ///< Allows barriers in conditional control flow. u64 device_access_memory{}; ///< Total size of device local memory in bytes. u32 sets_per_pool{}; ///< Sets per Description Pool |