From eb67a45ca82bc01ac843c853fd3c17f2a90e0250 Mon Sep 17 00:00:00 2001 From: ameerj Date: Mon, 26 Oct 2020 23:07:36 -0400 Subject: video_core: NVDEC Implementation This commit aims to implement the NVDEC (Nvidia Decoder) functionality, with video frame decoding being handled by the FFmpeg library. The process begins with Ioctl commands being sent to the NVDEC and VIC (Video Image Composer) emulated devices. These allocate the necessary GPU buffers for the frame data, along with providing information on the incoming video data. A Submit command then signals the GPU to process and decode the frame data. To decode the frame, the respective codec's header must be manually composed from the information provided by NVDEC, then sent with the raw frame data to the ffmpeg library. Currently, H264 and VP9 are supported, with VP9 having some minor artifacting issues related mainly to the reference frame composition in its uncompressed header. Async GPU is not properly implemented at the moment. Co-Authored-By: David <25727384+ogniK5377@users.noreply.github.com> --- src/video_core/gpu_thread.h | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-) (limited to 'src/video_core/gpu_thread.h') diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h index 5a28335d6..32a34e3a7 100644 --- a/src/video_core/gpu_thread.h +++ b/src/video_core/gpu_thread.h @@ -37,6 +37,14 @@ struct SubmitListCommand final { Tegra::CommandList entries; }; +/// Command to signal to the GPU thread that a cdma command list is ready for processing +struct SubmitChCommandEntries final { + explicit SubmitChCommandEntries(Tegra::ChCommandHeaderList&& entries) + : entries{std::move(entries)} {} + + Tegra::ChCommandHeaderList entries; +}; + /// Command to signal to the GPU thread that a swap buffers is pending struct SwapBuffersCommand final { explicit SwapBuffersCommand(std::optional framebuffer) @@ -77,9 +85,9 @@ struct OnCommandListEndCommand final {}; struct GPUTickCommand final {}; using CommandData = - std::variant; + std::variant; struct CommandDataContainer { CommandDataContainer() = default; @@ -109,11 +117,14 @@ public: /// Creates and starts the GPU thread. void StartThread(VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context, - Tegra::DmaPusher& dma_pusher); + Tegra::DmaPusher& dma_pusher, Tegra::CDmaPusher& cdma_pusher); /// Push GPU command entries to be processed void SubmitList(Tegra::CommandList&& entries); + /// Push GPU CDMA command buffer entries to be processed + void SubmitCommandBuffer(Tegra::ChCommandHeaderList&& entries); + /// Swap buffers (render frame) void SwapBuffers(const Tegra::FramebufferConfig* framebuffer); -- cgit v1.2.3 From 677a8b208d47d0d2397197ce74c7039a8ea79d20 Mon Sep 17 00:00:00 2001 From: Lioncash Date: Fri, 4 Dec 2020 14:39:12 -0500 Subject: video_core: Resolve more variable shadowing scenarios Resolves variable shadowing scenarios up to the end of the OpenGL code to make it nicer to review. The rest will be resolved in a following commit. --- src/video_core/gpu_thread.h | 25 ++++++++++++------------- 1 file changed, 12 insertions(+), 13 deletions(-) (limited to 'src/video_core/gpu_thread.h') diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h index 32a34e3a7..f1c52cd9e 100644 --- a/src/video_core/gpu_thread.h +++ b/src/video_core/gpu_thread.h @@ -32,30 +32,30 @@ struct EndProcessingCommand final {}; /// Command to signal to the GPU thread that a command list is ready for processing struct SubmitListCommand final { - explicit SubmitListCommand(Tegra::CommandList&& entries) : entries{std::move(entries)} {} + explicit SubmitListCommand(Tegra::CommandList&& entries_) : entries{std::move(entries_)} {} Tegra::CommandList entries; }; /// Command to signal to the GPU thread that a cdma command list is ready for processing struct SubmitChCommandEntries final { - explicit SubmitChCommandEntries(Tegra::ChCommandHeaderList&& entries) - : entries{std::move(entries)} {} + explicit SubmitChCommandEntries(Tegra::ChCommandHeaderList&& entries_) + : entries{std::move(entries_)} {} Tegra::ChCommandHeaderList entries; }; /// Command to signal to the GPU thread that a swap buffers is pending struct SwapBuffersCommand final { - explicit SwapBuffersCommand(std::optional framebuffer) - : framebuffer{std::move(framebuffer)} {} + explicit SwapBuffersCommand(std::optional framebuffer_) + : framebuffer{std::move(framebuffer_)} {} std::optional framebuffer; }; /// Command to signal to the GPU thread to flush a region struct FlushRegionCommand final { - explicit constexpr FlushRegionCommand(VAddr addr, u64 size) : addr{addr}, size{size} {} + explicit constexpr FlushRegionCommand(VAddr addr_, u64 size_) : addr{addr_}, size{size_} {} VAddr addr; u64 size; @@ -63,7 +63,7 @@ struct FlushRegionCommand final { /// Command to signal to the GPU thread to invalidate a region struct InvalidateRegionCommand final { - explicit constexpr InvalidateRegionCommand(VAddr addr, u64 size) : addr{addr}, size{size} {} + explicit constexpr InvalidateRegionCommand(VAddr addr_, u64 size_) : addr{addr_}, size{size_} {} VAddr addr; u64 size; @@ -71,8 +71,8 @@ struct InvalidateRegionCommand final { /// Command to signal to the GPU thread to flush and invalidate a region struct FlushAndInvalidateRegionCommand final { - explicit constexpr FlushAndInvalidateRegionCommand(VAddr addr, u64 size) - : addr{addr}, size{size} {} + explicit constexpr FlushAndInvalidateRegionCommand(VAddr addr_, u64 size_) + : addr{addr_}, size{size_} {} VAddr addr; u64 size; @@ -92,8 +92,8 @@ using CommandData = struct CommandDataContainer { CommandDataContainer() = default; - CommandDataContainer(CommandData&& data, u64 next_fence) - : data{std::move(data)}, fence{next_fence} {} + explicit CommandDataContainer(CommandData&& data_, u64 next_fence_) + : data{std::move(data_)}, fence{next_fence_} {} CommandData data; u64 fence{}; @@ -112,7 +112,7 @@ struct SynchState final { /// Class used to manage the GPU thread class ThreadManager final { public: - explicit ThreadManager(Core::System& system); + explicit ThreadManager(Core::System& system_); ~ThreadManager(); /// Creates and starts the GPU thread. @@ -146,7 +146,6 @@ private: /// Pushes a command to be executed by the GPU thread u64 PushCommand(CommandData&& command_data); -private: SynchState state; Core::System& system; std::thread thread; -- cgit v1.2.3 From 14c825bd1c37b2444e858bf1a75fb77455b4eb52 Mon Sep 17 00:00:00 2001 From: bunnei Date: Fri, 11 Dec 2020 22:26:14 -0800 Subject: video_core: gpu: Refactor out synchronous/asynchronous GPU implementations. - We must always use a GPU thread now, even with synchronous GPU. --- src/video_core/gpu_thread.h | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'src/video_core/gpu_thread.h') diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h index f1c52cd9e..0071195d6 100644 --- a/src/video_core/gpu_thread.h +++ b/src/video_core/gpu_thread.h @@ -10,8 +10,9 @@ #include #include #include + #include "common/threadsafe_queue.h" -#include "video_core/gpu.h" +#include "video_core/framebuffer_config.h" namespace Tegra { struct FramebufferConfig; @@ -25,6 +26,10 @@ class GraphicsContext; class System; } // namespace Core +namespace VideoCore { + class RendererBase; +} // namespace VideoCore + namespace VideoCommon::GPUThread { /// Command to signal to the GPU thread that processing has ended -- cgit v1.2.3 From 40571c073faa02a6a4301e7f0ce365ef50a400aa Mon Sep 17 00:00:00 2001 From: bunnei Date: Sat, 12 Dec 2020 00:24:33 -0800 Subject: video_core: gpu: Implement synchronous mode using threaded GPU. --- src/video_core/gpu_thread.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'src/video_core/gpu_thread.h') diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h index 0071195d6..2775629e7 100644 --- a/src/video_core/gpu_thread.h +++ b/src/video_core/gpu_thread.h @@ -27,7 +27,7 @@ class System; } // namespace Core namespace VideoCore { - class RendererBase; +class RendererBase; } // namespace VideoCore namespace VideoCommon::GPUThread { @@ -117,7 +117,7 @@ struct SynchState final { /// Class used to manage the GPU thread class ThreadManager final { public: - explicit ThreadManager(Core::System& system_); + explicit ThreadManager(Core::System& system_, bool is_async_); ~ThreadManager(); /// Creates and starts the GPU thread. @@ -155,6 +155,7 @@ private: Core::System& system; std::thread thread; std::thread::id thread_id; + const bool is_async; }; } // namespace VideoCommon::GPUThread -- cgit v1.2.3