yuzu/src/video_core/renderer_vulkan/vk_scheduler.h

// SPDX-FileCopyrightText: Copyright 2019 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later

#pragma once

#include <condition_variable>
#include <cstddef>
#include <memory>
#include <thread>
#include <utility>
#include <queue>

#include "common/alignment.h"
#include "common/common_types.h"
#include "video_core/renderer_vulkan/vk_master_semaphore.h"
#include "video_core/vulkan_common/vulkan_wrapper.h"

namespace Vulkan {

class CommandPool;
class Device;
class Framebuffer;
class GraphicsPipeline;
class StateTracker;
class VKQueryCache;

/// The scheduler abstracts command buffer and fence management with an interface that's able to do
/// OpenGL-like operations on Vulkan command buffers.
class VKScheduler {
public:
    explicit VKScheduler(const Device& device, StateTracker& state_tracker);
    ~VKScheduler();

    /// Sends the current execution context to the GPU.
    void Flush(VkSemaphore signal_semaphore = nullptr, VkSemaphore wait_semaphore = nullptr);

    /// Sends the current execution context to the GPU and waits for it to complete.
    void Finish(VkSemaphore signal_semaphore = nullptr, VkSemaphore wait_semaphore = nullptr);

    /// Waits for the worker thread to finish executing everything. After this function returns it's
    /// safe to touch worker resources.
    void WaitWorker();

    /// Sends currently recorded work to the worker thread.
    void DispatchWork();

    /// Requests to begin a renderpass.
    void RequestRenderpass(const Framebuffer* framebuffer);

    /// Requests the current executino context to be able to execute operations only allowed outside
    /// of a renderpass.
    void RequestOutsideRenderPassOperationContext();

    /// Update the pipeline to the current execution context.
    bool UpdateGraphicsPipeline(GraphicsPipeline* pipeline);

    /// Update the rescaling state. Returns true if the state has to be updated.
    bool UpdateRescaling(bool is_rescaling);

    /// Invalidates current command buffer state except for render passes
    void InvalidateState();

    /// Assigns the query cache.
    void SetQueryCache(VKQueryCache& query_cache_) {
        query_cache = &query_cache_;
    }

    /// Send work to a separate thread.
    template <typename T>
    void Record(T&& command) {
        if (chunk->Record(command)) {
            return;
        }
        DispatchWork();
        (void)chunk->Record(command);
    }

    /// Returns the current command buffer tick.
    [[nodiscard]] u64 CurrentTick() const noexcept {
        return master_semaphore->CurrentTick();
    }

    /// Returns true when a tick has been triggered by the GPU.
    [[nodiscard]] bool IsFree(u64 tick) const noexcept {
        return master_semaphore->IsFree(tick);
    }

    /// Waits for the given tick to trigger on the GPU.
    void Wait(u64 tick) {
        if (tick >= master_semaphore->CurrentTick()) {
            // Make sure we are not waiting for the current tick without signalling
            Flush();
        }
        master_semaphore->Wait(tick);
    }

    /// Returns the master timeline semaphore.
    [[nodiscard]] MasterSemaphore& GetMasterSemaphore() const noexcept {
        return *master_semaphore;
    }

private:
    class Command {
    public:
        virtual ~Command() = default;

        virtual void Execute(vk::CommandBuffer cmdbuf) const = 0;

        Command* GetNext() const {
            return next;
        }

        void SetNext(Command* next_) {
            next = next_;
        }

    private:
        Command* next = nullptr;
    };

    template <typename T>
    class TypedCommand final : public Command {
    public:
        explicit TypedCommand(T&& command_) : command{std::move(command_)} {}
        ~TypedCommand() override = default;

        TypedCommand(TypedCommand&&) = delete;
        TypedCommand& operator=(TypedCommand&&) = delete;

        void Execute(vk::CommandBuffer cmdbuf) const override {
            command(cmdbuf);
        }

    private:
        T command;
    };

    class CommandChunk final {
    public:
        void ExecuteAll(vk::CommandBuffer cmdbuf);

        template <typename T>
        bool Record(T& command) {
            using FuncType = TypedCommand<T>;
            static_assert(sizeof(FuncType) < sizeof(data), "Lambda is too large");

            recorded_counts++;
            command_offset = Common::AlignUp(command_offset, alignof(FuncType));
            if (command_offset > sizeof(data) - sizeof(FuncType)) {
                return false;
            }
            Command* const current_last = last;
            last = new (data.data() + command_offset) FuncType(std::move(command));

            if (current_last) {
                current_last->SetNext(last);
            } else {
                first = last;
            }
            command_offset += sizeof(FuncType);
            return true;
        }

        void MarkSubmit() {
            submit = true;
        }

        bool Empty() const {
            return recorded_counts == 0;
        }

        bool HasSubmit() const {
            return submit;
        }

    private:
        Command* first = nullptr;
        Command* last = nullptr;

        size_t recorded_counts = 0;
        size_t command_offset = 0;
        bool submit = false;
        alignas(std::max_align_t) std::array<u8, 0x8000> data{};
    };

    struct State {
        VkRenderPass renderpass = nullptr;
        VkFramebuffer framebuffer = nullptr;
        VkExtent2D render_area = {0, 0};
        GraphicsPipeline* graphics_pipeline = nullptr;
        bool is_rescaling = false;
        bool rescaling_defined = false;
    };

    void WorkerThread(std::stop_token stop_token);

    void AllocateWorkerCommandBuffer();

    void SubmitExecution(VkSemaphore signal_semaphore, VkSemaphore wait_semaphore);

    void AllocateNewContext();

    void EndPendingOperations();

    void EndRenderPass();

    void AcquireNewChunk();

    const Device& device;
    StateTracker& state_tracker;

    std::unique_ptr<MasterSemaphore> master_semaphore;
    std::unique_ptr<CommandPool> command_pool;

    VKQueryCache* query_cache = nullptr;

    vk::CommandBuffer current_cmdbuf;

    std::unique_ptr<CommandChunk> chunk;

    State state;

    u32 num_renderpass_images = 0;
    std::array<VkImage, 9> renderpass_images{};
    std::array<VkImageSubresourceRange, 9> renderpass_image_ranges{};

    std::queue<std::unique_ptr<CommandChunk>> work_queue;
    std::vector<std::unique_ptr<CommandChunk>> chunk_reserve;
    std::mutex reserve_mutex;
    std::mutex work_mutex;
    std::condition_variable_any work_cv;
    std::condition_variable wait_cv;
    std::jthread worker_thread;
};

} // namespace Vulkan