dynarmic: Update and enable DYNARMIC_IGNORE_ASSERTS

decoders: Optimize memcpy for the other functions

externals/CMakeLists.txt

@@ -7,7 +7,9 @@ include(DownloadExternals)
 # xbyak
 if (ARCHITECTURE_x86 OR ARCHITECTURE_x86_64)
     add_library(xbyak INTERFACE)
-    target_include_directories(xbyak SYSTEM INTERFACE ./xbyak/xbyak)
+    file(MAKE_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/xbyak/include)
+    file(COPY ${CMAKE_CURRENT_SOURCE_DIR}/xbyak/xbyak DESTINATION ${CMAKE_CURRENT_BINARY_DIR}/xbyak/include)
+    target_include_directories(xbyak SYSTEM INTERFACE ${CMAKE_CURRENT_BINARY_DIR}/xbyak/include)
     target_compile_definitions(xbyak INTERFACE XBYAK_NO_OP_NAMES)
 endif()
 
@@ -19,6 +21,7 @@ target_include_directories(catch-single-include INTERFACE catch/single_include)
 if (ARCHITECTURE_x86_64)
     set(DYNARMIC_TESTS OFF)
     set(DYNARMIC_NO_BUNDLED_FMT ON)
+    set(DYNARMIC_IGNORE_ASSERTS ON CACHE BOOL "" FORCE)
     add_subdirectory(dynarmic)
 endif()
 

src/common/logging/backend.cpp

@@ -2,9 +2,13 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
 #include <atomic>
 #include <chrono>
 #include <climits>
+#include <condition_variable>
+#include <memory>
+#include <mutex>
 #include <thread>
 #include <vector>
 
@@ -12,173 +16,28 @@
 #include <windows.h> // For OutputDebugStringW
 #endif
 
+#include "common/assert.h"
 #include "common/fs/file.h"
 #include "common/fs/fs.h"
-#include "common/fs/fs_paths.h"
-#include "common/fs/path_util.h"
 #include "common/literals.h"
 
 #include "common/logging/backend.h"
 #include "common/logging/log.h"
 #include "common/logging/text_formatter.h"
 #include "common/settings.h"
-#ifdef _WIN32
 #include "common/string_util.h"
-#endif
 #include "common/threadsafe_queue.h"
 
 namespace Common::Log {
 
-namespace {
-
-/**
- * Interface for logging backends.
- */
-class Backend {
-public:
-    virtual ~Backend() = default;
-
-    virtual void Write(const Entry& entry) = 0;
-
-    virtual void EnableForStacktrace() = 0;
-
-    virtual void Flush() = 0;
-};
-
-/**
- * Backend that writes to stderr and with color
- */
-class ColorConsoleBackend final : public Backend {
-public:
-    explicit ColorConsoleBackend() = default;
-
-    ~ColorConsoleBackend() override = default;
-
-    void Write(const Entry& entry) override {
-        if (enabled.load(std::memory_order_relaxed)) {
-            PrintColoredMessage(entry);
-        }
-    }
-
-    void Flush() override {
-        // stderr shouldn't be buffered
-    }
-
-    void EnableForStacktrace() override {
-        enabled = true;
-    }
-
-    void SetEnabled(bool enabled_) {
-        enabled = enabled_;
-    }
-
-private:
-    std::atomic_bool enabled{false};
-};
-
-/**
- * Backend that writes to a file passed into the constructor
- */
-class FileBackend final : public Backend {
-public:
-    explicit FileBackend(const std::filesystem::path& filename) {
-        auto old_filename = filename;
-        old_filename += ".old.txt";
-
-        // Existence checks are done within the functions themselves.
-        // We don't particularly care if these succeed or not.
-        static_cast<void>(FS::RemoveFile(old_filename));
-        static_cast<void>(FS::RenameFile(filename, old_filename));
-
-        file = std::make_unique<FS::IOFile>(filename, FS::FileAccessMode::Write,
-                                            FS::FileType::TextFile);
-    }
-
-    ~FileBackend() override = default;
-
-    void Write(const Entry& entry) override {
-        if (!enabled) {
-            return;
-        }
-
-        bytes_written += file->WriteString(FormatLogMessage(entry).append(1, '\n'));
-
-        using namespace Common::Literals;
-        // Prevent logs from exceeding a set maximum size in the event that log entries are spammed.
-        const auto write_limit = Settings::values.extended_logging ? 1_GiB : 100_MiB;
-        const bool write_limit_exceeded = bytes_written > write_limit;
-        if (entry.log_level >= Level::Error || write_limit_exceeded) {
-            if (write_limit_exceeded) {
-                // Stop writing after the write limit is exceeded.
-                // Don't close the file so we can print a stacktrace if necessary
-                enabled = false;
-            }
-            file->Flush();
-        }
-    }
-
-    void Flush() override {
-        file->Flush();
-    }
-
-    void EnableForStacktrace() override {
-        enabled = true;
-        bytes_written = 0;
-    }
-
-private:
-    std::unique_ptr<FS::IOFile> file;
-    bool enabled = true;
-    std::size_t bytes_written = 0;
-};
-
-/**
- * Backend that writes to Visual Studio's output window
- */
-class DebuggerBackend final : public Backend {
-public:
-    explicit DebuggerBackend() = default;
-
-    ~DebuggerBackend() override = default;
-
-    void Write(const Entry& entry) override {
-#ifdef _WIN32
-        ::OutputDebugStringW(UTF8ToUTF16W(FormatLogMessage(entry).append(1, '\n')).c_str());
-#endif
-    }
-
-    void Flush() override {}
-
-    void EnableForStacktrace() override {}
-};
-
-bool initialization_in_progress_suppress_logging = false;
-
 /**
  * Static state as a singleton.
  */
 class Impl {
 public:
     static Impl& Instance() {
-        if (!instance) {
-            abort();
-        }
-        return *instance;
-    }
-
-    static void Initialize() {
-        if (instance) {
-            abort();
-        }
-        using namespace Common::FS;
-        initialization_in_progress_suppress_logging = true;
-        const auto& log_dir = GetYuzuPath(YuzuPath::LogDir);
-        void(CreateDir(log_dir));
-        Filter filter;
-        filter.ParseFilterString(Settings::values.log_filter.GetValue());
-        instance = std::unique_ptr<Impl, decltype(&Deleter)>(new Impl(log_dir / LOG_FILE, filter),
-                                                             Deleter);
-        initialization_in_progress_suppress_logging = false;
+        static Impl backend;
+        return backend;
     }
 
     Impl(const Impl&) = delete;
@@ -187,54 +46,74 @@ public:
     Impl(Impl&&) = delete;
     Impl& operator=(Impl&&) = delete;
 
+    void PushEntry(Class log_class, Level log_level, const char* filename, unsigned int line_num,
+                   const char* function, std::string message) {
+        message_queue.Push(
+            CreateEntry(log_class, log_level, filename, line_num, function, std::move(message)));
+    }
+
+    void AddBackend(std::unique_ptr<Backend> backend) {
+        std::lock_guard lock{writing_mutex};
+        backends.push_back(std::move(backend));
+    }
+
+    void RemoveBackend(std::string_view backend_name) {
+        std::lock_guard lock{writing_mutex};
+
+        std::erase_if(backends, [&backend_name](const auto& backend) {
+            return backend_name == backend->GetName();
+        });
+    }
+
+    const Filter& GetGlobalFilter() const {
+        return filter;
+    }
+
     void SetGlobalFilter(const Filter& f) {
         filter = f;
     }
 
-    void SetColorConsoleBackendEnabled(bool enabled) {
-        color_console_backend.SetEnabled(enabled);
-    }
-
-    void PushEntry(Class log_class, Level log_level, const char* filename, unsigned int line_num,
-                   const char* function, std::string message) {
-        if (!filter.CheckMessage(log_class, log_level))
-            return;
-        const Entry& entry =
-            CreateEntry(log_class, log_level, filename, line_num, function, std::move(message));
-        message_queue.Push(entry);
+    Backend* GetBackend(std::string_view backend_name) {
+        const auto it =
+            std::find_if(backends.begin(), backends.end(),
+                         [&backend_name](const auto& i) { return backend_name == i->GetName(); });
+        if (it == backends.end())
+            return nullptr;
+        return it->get();
     }
 
 private:
-    Impl(const std::filesystem::path& file_backend_filename, const Filter& filter_)
-        : filter{filter_}, file_backend{file_backend_filename}, backend_thread{std::thread([this] {
-              Common::SetCurrentThreadName("yuzu:Log");
-              Entry entry;
-              const auto write_logs = [this, &entry]() {
-                  ForEachBackend([&entry](Backend& backend) { backend.Write(entry); });
-              };
-              while (true) {
-                  entry = message_queue.PopWait();
-                  if (entry.final_entry) {
-                      break;
-                  }
-                  write_logs();
-              }
-              // Drain the logging queue. Only writes out up to MAX_LOGS_TO_WRITE to prevent a
-              // case where a system is repeatedly spamming logs even on close.
-              int max_logs_to_write = filter.IsDebug() ? INT_MAX : 100;
-              while (max_logs_to_write-- && message_queue.Pop(entry)) {
-                  write_logs();
-              }
-          })} {}
+    Impl() {
+        backend_thread = std::thread([&] {
+            Entry entry;
+            auto write_logs = [&](Entry& e) {
+                std::lock_guard lock{writing_mutex};
+                for (const auto& backend : backends) {
+                    backend->Write(e);
+                }
+            };
+            while (true) {
+                entry = message_queue.PopWait();
+                if (entry.final_entry) {
+                    break;
+                }
+                write_logs(entry);
+            }
 
-    ~Impl() {
-        StopBackendThread();
+            // Drain the logging queue. Only writes out up to MAX_LOGS_TO_WRITE to prevent a
+            // case where a system is repeatedly spamming logs even on close.
+            const int MAX_LOGS_TO_WRITE = filter.IsDebug() ? INT_MAX : 100;
+            int logs_written = 0;
+            while (logs_written++ < MAX_LOGS_TO_WRITE && message_queue.Pop(entry)) {
+                write_logs(entry);
+            }
+        });
     }
 
-    void StopBackendThread() {
-        Entry stop_entry{};
-        stop_entry.final_entry = true;
-        message_queue.Push(stop_entry);
+    ~Impl() {
+        Entry entry;
+        entry.final_entry = true;
+        message_queue.Push(entry);
         backend_thread.join();
     }
 
@@ -256,51 +135,100 @@ private:
         };
     }
 
-    void ForEachBackend(auto lambda) {
-        lambda(static_cast<Backend&>(debugger_backend));
-        lambda(static_cast<Backend&>(color_console_backend));
-        lambda(static_cast<Backend&>(file_backend));
-    }
-
-    static void Deleter(Impl* ptr) {
-        delete ptr;
-    }
-
-    static inline std::unique_ptr<Impl, decltype(&Deleter)> instance{nullptr, Deleter};
-
-    Filter filter;
-    DebuggerBackend debugger_backend{};
-    ColorConsoleBackend color_console_backend{};
-    FileBackend file_backend;
-
+    std::mutex writing_mutex;
     std::thread backend_thread;
-    MPSCQueue<Entry> message_queue{};
+    std::vector<std::unique_ptr<Backend>> backends;
+    MPSCQueue<Entry> message_queue;
+    Filter filter;
     std::chrono::steady_clock::time_point time_origin{std::chrono::steady_clock::now()};
 };
-} // namespace
 
-void Initialize() {
-    Impl::Initialize();
+ConsoleBackend::~ConsoleBackend() = default;
+
+void ConsoleBackend::Write(const Entry& entry) {
+    PrintMessage(entry);
 }
 
-void DisableLoggingInTests() {
-    initialization_in_progress_suppress_logging = true;
+ColorConsoleBackend::~ColorConsoleBackend() = default;
+
+void ColorConsoleBackend::Write(const Entry& entry) {
+    PrintColoredMessage(entry);
+}
+
+FileBackend::FileBackend(const std::filesystem::path& filename) {
+    auto old_filename = filename;
+    old_filename += ".old.txt";
+
+    // Existence checks are done within the functions themselves.
+    // We don't particularly care if these succeed or not.
+    FS::RemoveFile(old_filename);
+    void(FS::RenameFile(filename, old_filename));
+
+    file =
+        std::make_unique<FS::IOFile>(filename, FS::FileAccessMode::Write, FS::FileType::TextFile);
+}
+
+FileBackend::~FileBackend() = default;
+
+void FileBackend::Write(const Entry& entry) {
+    if (!file->IsOpen()) {
+        return;
+    }
+
+    using namespace Common::Literals;
+    // Prevent logs from exceeding a set maximum size in the event that log entries are spammed.
+    constexpr std::size_t MAX_BYTES_WRITTEN = 100_MiB;
+    constexpr std::size_t MAX_BYTES_WRITTEN_EXTENDED = 1_GiB;
+
+    const bool write_limit_exceeded =
+        bytes_written > MAX_BYTES_WRITTEN_EXTENDED ||
+        (bytes_written > MAX_BYTES_WRITTEN && !Settings::values.extended_logging);
+
+    // Close the file after the write limit is exceeded.
+    if (write_limit_exceeded) {
+        file->Close();
+        return;
+    }
+
+    bytes_written += file->WriteString(FormatLogMessage(entry).append(1, '\n'));
+    if (entry.log_level >= Level::Error) {
+        file->Flush();
+    }
+}
+
+DebuggerBackend::~DebuggerBackend() = default;
+
+void DebuggerBackend::Write(const Entry& entry) {
+#ifdef _WIN32
+    ::OutputDebugStringW(UTF8ToUTF16W(FormatLogMessage(entry).append(1, '\n')).c_str());
+#endif
 }
 
 void SetGlobalFilter(const Filter& filter) {
     Impl::Instance().SetGlobalFilter(filter);
 }
 
-void SetColorConsoleBackendEnabled(bool enabled) {
-    Impl::Instance().SetColorConsoleBackendEnabled(enabled);
+void AddBackend(std::unique_ptr<Backend> backend) {
+    Impl::Instance().AddBackend(std::move(backend));
+}
+
+void RemoveBackend(std::string_view backend_name) {
+    Impl::Instance().RemoveBackend(backend_name);
+}
+
+Backend* GetBackend(std::string_view backend_name) {
+    return Impl::Instance().GetBackend(backend_name);
 }
 
 void FmtLogMessageImpl(Class log_class, Level log_level, const char* filename,
                        unsigned int line_num, const char* function, const char* format,
                        const fmt::format_args& args) {
-    if (!initialization_in_progress_suppress_logging) {
-        Impl::Instance().PushEntry(log_class, log_level, filename, line_num, function,
-                                   fmt::vformat(format, args));
-    }
+    auto& instance = Impl::Instance();
+    const auto& filter = instance.GetGlobalFilter();
+    if (!filter.CheckMessage(log_class, log_level))
+        return;
+
+    instance.PushEntry(log_class, log_level, filename, line_num, function,
+                       fmt::vformat(format, args));
 }
 } // namespace Common::Log

src/common/logging/backend.h

@@ -5,21 +5,120 @@
 #pragma once
 
 #include <filesystem>
+#include <memory>
+#include <string>
+#include <string_view>
 #include "common/logging/filter.h"
+#include "common/logging/log.h"
+
+namespace Common::FS {
+class IOFile;
+}
 
 namespace Common::Log {
 
 class Filter;
 
-/// Initializes the logging system. This should be the first thing called in main.
-void Initialize();
+/**
+ * Interface for logging backends. As loggers can be created and removed at runtime, this can be
+ * used by a frontend for adding a custom logging backend as needed
+ */
+class Backend {
+public:
+    virtual ~Backend() = default;
 
-void DisableLoggingInTests();
+    virtual void SetFilter(const Filter& new_filter) {
+        filter = new_filter;
+    }
+    virtual const char* GetName() const = 0;
+    virtual void Write(const Entry& entry) = 0;
+
+private:
+    Filter filter;
+};
 
 /**
- * The global filter will prevent any messages from even being processed if they are filtered.
+ * Backend that writes to stderr without any color commands
+ */
+class ConsoleBackend : public Backend {
+public:
+    ~ConsoleBackend() override;
+
+    static const char* Name() {
+        return "console";
+    }
+    const char* GetName() const override {
+        return Name();
+    }
+    void Write(const Entry& entry) override;
+};
+
+/**
+ * Backend that writes to stderr and with color
+ */
+class ColorConsoleBackend : public Backend {
+public:
+    ~ColorConsoleBackend() override;
+
+    static const char* Name() {
+        return "color_console";
+    }
+
+    const char* GetName() const override {
+        return Name();
+    }
+    void Write(const Entry& entry) override;
+};
+
+/**
+ * Backend that writes to a file passed into the constructor
+ */
+class FileBackend : public Backend {
+public:
+    explicit FileBackend(const std::filesystem::path& filename);
+    ~FileBackend() override;
+
+    static const char* Name() {
+        return "file";
+    }
+
+    const char* GetName() const override {
+        return Name();
+    }
+
+    void Write(const Entry& entry) override;
+
+private:
+    std::unique_ptr<FS::IOFile> file;
+    std::size_t bytes_written = 0;
+};
+
+/**
+ * Backend that writes to Visual Studio's output window
+ */
+class DebuggerBackend : public Backend {
+public:
+    ~DebuggerBackend() override;
+
+    static const char* Name() {
+        return "debugger";
+    }
+    const char* GetName() const override {
+        return Name();
+    }
+    void Write(const Entry& entry) override;
+};
+
+void AddBackend(std::unique_ptr<Backend> backend);
+
+void RemoveBackend(std::string_view backend_name);
+
+Backend* GetBackend(std::string_view backend_name);
+
+/**
+ * The global filter will prevent any messages from even being processed if they are filtered. Each
+ * backend can have a filter, but if the level is lower than the global filter, the backend will
+ * never get the message
  */
 void SetGlobalFilter(const Filter& filter);
-
-void SetColorConsoleBackendEnabled(bool enabled);
-} // namespace Common::Log
+} // namespace Common::Log

src/common/threadsafe_queue.h

@@ -46,15 +46,13 @@ public:
         ElementPtr* new_ptr = new ElementPtr();
         write_ptr->next.store(new_ptr, std::memory_order_release);
         write_ptr = new_ptr;
+        ++size;
 
-        const size_t previous_size{size++};
-
-        // Acquire the mutex and then immediately release it as a fence.
+        // cv_mutex must be held or else there will be a missed wakeup if the other thread is in the
+        // line before cv.wait
         // TODO(bunnei): This can be replaced with C++20 waitable atomics when properly supported.
         // See discussion on https://github.com/yuzu-emu/yuzu/pull/3173 for details.
-        if (previous_size == 0) {
-            std::lock_guard lock{cv_mutex};
-        }
+        std::lock_guard lock{cv_mutex};
         cv.notify_one();
     }
 

src/common/x64/xbyak_abi.h

@@ -6,7 +6,7 @@
 
 #include <bitset>
 #include <initializer_list>
-#include <xbyak.h>
+#include <xbyak/xbyak.h>
 #include "common/assert.h"
 
 namespace Common::X64 {

src/common/x64/xbyak_util.h

@@ -5,7 +5,7 @@
 #pragma once
 
 #include <type_traits>
-#include <xbyak.h>
+#include <xbyak/xbyak.h>
 #include "common/x64/xbyak_abi.h"
 
 namespace Common::X64 {

src/core/core.cpp

@@ -84,6 +84,8 @@ FileSys::StorageId GetStorageIdForFrontendSlot(
 
 } // Anonymous namespace
 
+/*static*/ System System::s_instance;
+
 FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
                                          const std::string& path) {
     // To account for split 00+01+etc files.
@@ -423,13 +425,6 @@ struct System::Impl {
 System::System() : impl{std::make_unique<Impl>(*this)} {}
 System::~System() = default;
 
-void System::InitializeGlobalInstance() {
-    if (s_instance) {
-        abort();
-    }
-    s_instance = std::unique_ptr<System>(new System);
-}
-
 CpuManager& System::GetCpuManager() {
     return impl->cpu_manager;
 }

src/core/core.h

@@ -121,14 +121,9 @@ public:
      * @returns Reference to the instance of the System singleton class.
      */
     [[deprecated("Use of the global system instance is deprecated")]] static System& GetInstance() {
-        if (!s_instance) {
-            abort();
-        }
-        return *s_instance;
+        return s_instance;
     }
 
-    static void InitializeGlobalInstance();
-
     /// Enumeration representing the return values of the System Initialize and Load process.
     enum class ResultStatus : u32 {
         Success,             ///< Succeeded
@@ -401,7 +396,7 @@ private:
     struct Impl;
     std::unique_ptr<Impl> impl;
 
-    inline static std::unique_ptr<System> s_instance{};
+    static System s_instance;
 };
 
 } // namespace Core

src/tests/common/param_package.cpp

@@ -4,13 +4,11 @@
 
 #include <catch2/catch.hpp>
 #include <math.h>
-#include "common/logging/backend.h"
 #include "common/param_package.h"
 
 namespace Common {
 
 TEST_CASE("ParamPackage", "[common]") {
-    Common::Log::DisableLoggingInTests();
     ParamPackage original{
         {"abc", "xyz"},
         {"def", "42"},

src/video_core/macro/macro_jit_x64.h

@@ -6,7 +6,7 @@
 
 #include <array>
 #include <bitset>
-#include <xbyak.h>
+#include <xbyak/xbyak.h>
 #include "common/bit_field.h"
 #include "common/common_types.h"
 #include "common/x64/xbyak_abi.h"

src/video_core/textures/decoders.cpp

@@ -84,34 +84,107 @@ template <bool TO_LINEAR>
 void Swizzle(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel, u32 width,
              u32 height, u32 depth, u32 block_height, u32 block_depth, u32 stride_alignment) {
     switch (bytes_per_pixel) {
-    case 1:
-        return SwizzleImpl<TO_LINEAR, 1>(output, input, width, height, depth, block_height,
+#define BPP_CASE(x)                                                                                \
+    case x:                                                                                        \
+        return SwizzleImpl<TO_LINEAR, x>(output, input, width, height, depth, block_height,        \
                                          block_depth, stride_alignment);
-    case 2:
-        return SwizzleImpl<TO_LINEAR, 2>(output, input, width, height, depth, block_height,
-                                         block_depth, stride_alignment);
-    case 3:
-        return SwizzleImpl<TO_LINEAR, 3>(output, input, width, height, depth, block_height,
-                                         block_depth, stride_alignment);
-    case 4:
-        return SwizzleImpl<TO_LINEAR, 4>(output, input, width, height, depth, block_height,
-                                         block_depth, stride_alignment);
-    case 6:
-        return SwizzleImpl<TO_LINEAR, 6>(output, input, width, height, depth, block_height,
-                                         block_depth, stride_alignment);
-    case 8:
-        return SwizzleImpl<TO_LINEAR, 8>(output, input, width, height, depth, block_height,
-                                         block_depth, stride_alignment);
-    case 12:
-        return SwizzleImpl<TO_LINEAR, 12>(output, input, width, height, depth, block_height,
-                                          block_depth, stride_alignment);
-    case 16:
-        return SwizzleImpl<TO_LINEAR, 16>(output, input, width, height, depth, block_height,
-                                          block_depth, stride_alignment);
+        BPP_CASE(1)
+        BPP_CASE(2)
+        BPP_CASE(3)
+        BPP_CASE(4)
+        BPP_CASE(6)
+        BPP_CASE(8)
+        BPP_CASE(12)
+        BPP_CASE(16)
+#undef BPP_CASE
     default:
         UNREACHABLE_MSG("Invalid bytes_per_pixel={}", bytes_per_pixel);
     }
 }
+
+template <u32 BYTES_PER_PIXEL>
+void SwizzleSubrect(u32 subrect_width, u32 subrect_height, u32 source_pitch, u32 swizzled_width,
+                    u8* swizzled_data, const u8* unswizzled_data, u32 block_height_bit,
+                    u32 offset_x, u32 offset_y) {
+    const u32 block_height = 1U << block_height_bit;
+    const u32 image_width_in_gobs =
+        (swizzled_width * BYTES_PER_PIXEL + (GOB_SIZE_X - 1)) / GOB_SIZE_X;
+    for (u32 line = 0; line < subrect_height; ++line) {
+        const u32 dst_y = line + offset_y;
+        const u32 gob_address_y =
+            (dst_y / (GOB_SIZE_Y * block_height)) * GOB_SIZE * block_height * image_width_in_gobs +
+            ((dst_y % (GOB_SIZE_Y * block_height)) / GOB_SIZE_Y) * GOB_SIZE;
+        const auto& table = SWIZZLE_TABLE[dst_y % GOB_SIZE_Y];
+        for (u32 x = 0; x < subrect_width; ++x) {
+            const u32 dst_x = x + offset_x;
+            const u32 gob_address =
+                gob_address_y + (dst_x * BYTES_PER_PIXEL / GOB_SIZE_X) * GOB_SIZE * block_height;
+            const u32 swizzled_offset = gob_address + table[(dst_x * BYTES_PER_PIXEL) % GOB_SIZE_X];
+            const u32 unswizzled_offset = line * source_pitch + x * BYTES_PER_PIXEL;
+
+            const u8* const source_line = unswizzled_data + unswizzled_offset;
+            u8* const dest_addr = swizzled_data + swizzled_offset;
+            std::memcpy(dest_addr, source_line, BYTES_PER_PIXEL);
+        }
+    }
+}
+
+template <u32 BYTES_PER_PIXEL>
+void UnswizzleSubrect(u32 line_length_in, u32 line_count, u32 pitch, u32 width, u32 block_height,
+                      u32 origin_x, u32 origin_y, u8* output, const u8* input) {
+    const u32 stride = width * BYTES_PER_PIXEL;
+    const u32 gobs_in_x = (stride + GOB_SIZE_X - 1) / GOB_SIZE_X;
+    const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block_height);
+
+    const u32 block_height_mask = (1U << block_height) - 1;
+    const u32 x_shift = GOB_SIZE_SHIFT + block_height;
+
+    for (u32 line = 0; line < line_count; ++line) {
+        const u32 src_y = line + origin_y;
+        const auto& table = SWIZZLE_TABLE[src_y % GOB_SIZE_Y];
+
+        const u32 block_y = src_y >> GOB_SIZE_Y_SHIFT;
+        const u32 src_offset_y = (block_y >> block_height) * block_size +
+                                 ((block_y & block_height_mask) << GOB_SIZE_SHIFT);
+        for (u32 column = 0; column < line_length_in; ++column) {
+            const u32 src_x = (column + origin_x) * BYTES_PER_PIXEL;
+            const u32 src_offset_x = (src_x >> GOB_SIZE_X_SHIFT) << x_shift;
+
+            const u32 swizzled_offset = src_offset_y + src_offset_x + table[src_x % GOB_SIZE_X];
+            const u32 unswizzled_offset = line * pitch + column * BYTES_PER_PIXEL;
+
+            std::memcpy(output + unswizzled_offset, input + swizzled_offset, BYTES_PER_PIXEL);
+        }
+    }
+}
+
+template <u32 BYTES_PER_PIXEL>
+void SwizzleSliceToVoxel(u32 line_length_in, u32 line_count, u32 pitch, u32 width, u32 height,
+                         u32 block_height, u32 block_depth, u32 origin_x, u32 origin_y, u8* output,
+                         const u8* input) {
+    UNIMPLEMENTED_IF(origin_x > 0);
+    UNIMPLEMENTED_IF(origin_y > 0);
+
+    const u32 stride = width * BYTES_PER_PIXEL;
+    const u32 gobs_in_x = (stride + GOB_SIZE_X - 1) / GOB_SIZE_X;
+    const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block_height + block_depth);
+
+    const u32 block_height_mask = (1U << block_height) - 1;
+    const u32 x_shift = static_cast<u32>(GOB_SIZE_SHIFT) + block_height + block_depth;
+
+    for (u32 line = 0; line < line_count; ++line) {
+        const auto& table = SWIZZLE_TABLE[line % GOB_SIZE_Y];
+        const u32 block_y = line / GOB_SIZE_Y;
+        const u32 dst_offset_y =
+            (block_y >> block_height) * block_size + (block_y & block_height_mask) * GOB_SIZE;
+        for (u32 x = 0; x < line_length_in; ++x) {
+            const u32 dst_offset =
+                ((x / GOB_SIZE_X) << x_shift) + dst_offset_y + table[x % GOB_SIZE_X];
+            const u32 src_offset = x * BYTES_PER_PIXEL + line * pitch;
+            std::memcpy(output + dst_offset, input + src_offset, BYTES_PER_PIXEL);
+        }
+    }
+}
 } // Anonymous namespace
 
 void UnswizzleTexture(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel,
@@ -131,81 +204,67 @@ void SwizzleTexture(std::span<u8> output, std::span<const u8> input, u32 bytes_p
 void SwizzleSubrect(u32 subrect_width, u32 subrect_height, u32 source_pitch, u32 swizzled_width,
                     u32 bytes_per_pixel, u8* swizzled_data, const u8* unswizzled_data,
                     u32 block_height_bit, u32 offset_x, u32 offset_y) {
-    const u32 block_height = 1U << block_height_bit;
-    const u32 image_width_in_gobs =
-        (swizzled_width * bytes_per_pixel + (GOB_SIZE_X - 1)) / GOB_SIZE_X;
-    for (u32 line = 0; line < subrect_height; ++line) {
-        const u32 dst_y = line + offset_y;
-        const u32 gob_address_y =
-            (dst_y / (GOB_SIZE_Y * block_height)) * GOB_SIZE * block_height * image_width_in_gobs +
-            ((dst_y % (GOB_SIZE_Y * block_height)) / GOB_SIZE_Y) * GOB_SIZE;
-        const auto& table = SWIZZLE_TABLE[dst_y % GOB_SIZE_Y];
-        for (u32 x = 0; x < subrect_width; ++x) {
-            const u32 dst_x = x + offset_x;
-            const u32 gob_address =
-                gob_address_y + (dst_x * bytes_per_pixel / GOB_SIZE_X) * GOB_SIZE * block_height;
-            const u32 swizzled_offset = gob_address + table[(dst_x * bytes_per_pixel) % GOB_SIZE_X];
-            const u32 unswizzled_offset = line * source_pitch + x * bytes_per_pixel;
-
-            const u8* const source_line = unswizzled_data + unswizzled_offset;
-            u8* const dest_addr = swizzled_data + swizzled_offset;
-            std::memcpy(dest_addr, source_line, bytes_per_pixel);
-        }
+    switch (bytes_per_pixel) {
+#define BPP_CASE(x)                                                                                \
+    case x:                                                                                        \
+        return SwizzleSubrect<x>(subrect_width, subrect_height, source_pitch, swizzled_width,      \
+                                 swizzled_data, unswizzled_data, block_height_bit, offset_x,       \
+                                 offset_y);
+        BPP_CASE(1)
+        BPP_CASE(2)
+        BPP_CASE(3)
+        BPP_CASE(4)
+        BPP_CASE(6)
+        BPP_CASE(8)
+        BPP_CASE(12)
+        BPP_CASE(16)
+#undef BPP_CASE
+    default:
+        UNREACHABLE_MSG("Invalid bytes_per_pixel={}", bytes_per_pixel);
     }
 }
 
 void UnswizzleSubrect(u32 line_length_in, u32 line_count, u32 pitch, u32 width, u32 bytes_per_pixel,
                       u32 block_height, u32 origin_x, u32 origin_y, u8* output, const u8* input) {
-    const u32 stride = width * bytes_per_pixel;
-    const u32 gobs_in_x = (stride + GOB_SIZE_X - 1) / GOB_SIZE_X;
-    const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block_height);
-
-    const u32 block_height_mask = (1U << block_height) - 1;
-    const u32 x_shift = GOB_SIZE_SHIFT + block_height;
-
-    for (u32 line = 0; line < line_count; ++line) {
-        const u32 src_y = line + origin_y;
-        const auto& table = SWIZZLE_TABLE[src_y % GOB_SIZE_Y];
-
-        const u32 block_y = src_y >> GOB_SIZE_Y_SHIFT;
-        const u32 src_offset_y = (block_y >> block_height) * block_size +
-                                 ((block_y & block_height_mask) << GOB_SIZE_SHIFT);
-        for (u32 column = 0; column < line_length_in; ++column) {
-            const u32 src_x = (column + origin_x) * bytes_per_pixel;
-            const u32 src_offset_x = (src_x >> GOB_SIZE_X_SHIFT) << x_shift;
-
-            const u32 swizzled_offset = src_offset_y + src_offset_x + table[src_x % GOB_SIZE_X];
-            const u32 unswizzled_offset = line * pitch + column * bytes_per_pixel;
-
-            std::memcpy(output + unswizzled_offset, input + swizzled_offset, bytes_per_pixel);
-        }
+    switch (bytes_per_pixel) {
+#define BPP_CASE(x)                                                                                \
+    case x:                                                                                        \
+        return UnswizzleSubrect<x>(line_length_in, line_count, pitch, width, block_height,         \
+                                   origin_x, origin_y, output, input);
+        BPP_CASE(1)
+        BPP_CASE(2)
+        BPP_CASE(3)
+        BPP_CASE(4)
+        BPP_CASE(6)
+        BPP_CASE(8)
+        BPP_CASE(12)
+        BPP_CASE(16)
+#undef BPP_CASE
+    default:
+        UNREACHABLE_MSG("Invalid bytes_per_pixel={}", bytes_per_pixel);
     }
 }
 
 void SwizzleSliceToVoxel(u32 line_length_in, u32 line_count, u32 pitch, u32 width, u32 height,
                          u32 bytes_per_pixel, u32 block_height, u32 block_depth, u32 origin_x,
                          u32 origin_y, u8* output, const u8* input) {
-    UNIMPLEMENTED_IF(origin_x > 0);
-    UNIMPLEMENTED_IF(origin_y > 0);
-
-    const u32 stride = width * bytes_per_pixel;
-    const u32 gobs_in_x = (stride + GOB_SIZE_X - 1) / GOB_SIZE_X;
-    const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block_height + block_depth);
-
-    const u32 block_height_mask = (1U << block_height) - 1;
-    const u32 x_shift = static_cast<u32>(GOB_SIZE_SHIFT) + block_height + block_depth;
-
-    for (u32 line = 0; line < line_count; ++line) {
-        const auto& table = SWIZZLE_TABLE[line % GOB_SIZE_Y];
-        const u32 block_y = line / GOB_SIZE_Y;
-        const u32 dst_offset_y =
-            (block_y >> block_height) * block_size + (block_y & block_height_mask) * GOB_SIZE;
-        for (u32 x = 0; x < line_length_in; ++x) {
-            const u32 dst_offset =
-                ((x / GOB_SIZE_X) << x_shift) + dst_offset_y + table[x % GOB_SIZE_X];
-            const u32 src_offset = x * bytes_per_pixel + line * pitch;
-            std::memcpy(output + dst_offset, input + src_offset, bytes_per_pixel);
-        }
+    switch (bytes_per_pixel) {
+#define BPP_CASE(x)                                                                                \
+    case x:                                                                                        \
+        return SwizzleSliceToVoxel<x>(line_length_in, line_count, pitch, width, height,            \
+                                      block_height, block_depth, origin_x, origin_y, output,       \
+                                      input);
+        BPP_CASE(1)
+        BPP_CASE(2)
+        BPP_CASE(3)
+        BPP_CASE(4)
+        BPP_CASE(6)
+        BPP_CASE(8)
+        BPP_CASE(12)
+        BPP_CASE(16)
+#undef BPP_CASE
+    default:
+        UNREACHABLE_MSG("Invalid bytes_per_pixel={}", bytes_per_pixel);
     }
 }
 
@@ -228,7 +287,7 @@ void SwizzleKepler(const u32 width, const u32 height, const u32 dst_x, const u32
             u8* dest_addr = swizzle_data + swizzled_offset;
             count++;
 
-            std::memcpy(dest_addr, source_line, 1);
+            *dest_addr = *source_line;
         }
     }
 }

src/yuzu/debugger/console.cpp

@@ -21,7 +21,6 @@ void ToggleConsole() {
         console_shown = UISettings::values.show_console.GetValue();
     }
 
-    using namespace Common::Log;
 #if defined(_WIN32) && !defined(_DEBUG)
     FILE* temp;
     if (UISettings::values.show_console) {
@@ -30,20 +29,24 @@ void ToggleConsole() {
             freopen_s(&temp, "CONIN$", "r", stdin);
             freopen_s(&temp, "CONOUT$", "w", stdout);
             freopen_s(&temp, "CONOUT$", "w", stderr);
-            SetColorConsoleBackendEnabled(true);
+            Common::Log::AddBackend(std::make_unique<Common::Log::ColorConsoleBackend>());
         }
     } else {
         if (FreeConsole()) {
             // In order to close the console, we have to also detach the streams on it.
             // Just redirect them to NUL if there is no console window
-            SetColorConsoleBackendEnabled(false);
+            Common::Log::RemoveBackend(Common::Log::ColorConsoleBackend::Name());
             freopen_s(&temp, "NUL", "r", stdin);
             freopen_s(&temp, "NUL", "w", stdout);
             freopen_s(&temp, "NUL", "w", stderr);
         }
     }
 #else
-    SetColorConsoleBackendEnabled(UISettings::values.show_console.GetValue());
+    if (UISettings::values.show_console) {
+        Common::Log::AddBackend(std::make_unique<Common::Log::ColorConsoleBackend>());
+    } else {
+        Common::Log::RemoveBackend(Common::Log::ColorConsoleBackend::Name());
+    }
 #endif
 }
 } // namespace Debugger

src/yuzu/main.cpp

@@ -175,6 +175,21 @@ void GMainWindow::ShowTelemetryCallout() {
 
 const int GMainWindow::max_recent_files_item;
 
+static void InitializeLogging() {
+    using namespace Common;
+
+    Log::Filter log_filter;
+    log_filter.ParseFilterString(Settings::values.log_filter.GetValue());
+    Log::SetGlobalFilter(log_filter);
+
+    const auto log_dir = FS::GetYuzuPath(FS::YuzuPath::LogDir);
+    void(FS::CreateDir(log_dir));
+    Log::AddBackend(std::make_unique<Log::FileBackend>(log_dir / LOG_FILE));
+#ifdef _WIN32
+    Log::AddBackend(std::make_unique<Log::DebuggerBackend>());
+#endif
+}
+
 static void RemoveCachedContents() {
     const auto cache_dir = Common::FS::GetYuzuPath(Common::FS::YuzuPath::CacheDir);
     const auto offline_fonts = cache_dir / "fonts";
@@ -192,6 +207,8 @@ GMainWindow::GMainWindow()
     : input_subsystem{std::make_shared<InputCommon::InputSubsystem>()},
       config{std::make_unique<Config>()}, vfs{std::make_shared<FileSys::RealVfsFilesystem>()},
       provider{std::make_unique<FileSys::ManualContentProvider>()} {
+    InitializeLogging();
+
     LoadTranslation();
 
     setAcceptDrops(true);
@@ -3381,7 +3398,6 @@ void GMainWindow::SetDiscordEnabled([[maybe_unused]] bool state) {
 #endif
 
 int main(int argc, char* argv[]) {
-    Common::Log::Initialize();
     Common::DetachedTasks detached_tasks;
     MicroProfileOnThreadCreate("Frontend");
     SCOPE_EXIT({ MicroProfileShutdown(); });
@@ -3421,7 +3437,6 @@ int main(int argc, char* argv[]) {
     // generating shaders
     setlocale(LC_ALL, "C");
 
-    Core::System::InitializeGlobalInstance();
     GMainWindow main_window;
     // After settings have been loaded by GMainWindow, apply the filter
     main_window.show();

src/yuzu_cmd/yuzu.cpp

@@ -74,14 +74,31 @@ static void PrintVersion() {
     std::cout << "yuzu " << Common::g_scm_branch << " " << Common::g_scm_desc << std::endl;
 }
 
+static void InitializeLogging() {
+    using namespace Common;
+
+    Log::Filter log_filter(Log::Level::Debug);
+    log_filter.ParseFilterString(static_cast<std::string>(Settings::values.log_filter));
+    Log::SetGlobalFilter(log_filter);
+
+    Log::AddBackend(std::make_unique<Log::ColorConsoleBackend>());
+
+    const auto& log_dir = FS::GetYuzuPath(FS::YuzuPath::LogDir);
+    void(FS::CreateDir(log_dir));
+    Log::AddBackend(std::make_unique<Log::FileBackend>(log_dir / LOG_FILE));
+#ifdef _WIN32
+    Log::AddBackend(std::make_unique<Log::DebuggerBackend>());
+#endif
+}
+
 /// Application entry point
 int main(int argc, char** argv) {
-    Common::Log::Initialize();
-    Common::Log::SetColorConsoleBackendEnabled(true);
     Common::DetachedTasks detached_tasks;
     Config config;
 
     int option_index = 0;
+
+    InitializeLogging();
 #ifdef _WIN32
     int argc_w;
     auto argv_w = CommandLineToArgvW(GetCommandLineW(), &argc_w);
@@ -146,7 +163,6 @@ int main(int argc, char** argv) {
         return -1;
     }
 
-    Core::System::InitializeGlobalInstance();
     auto& system{Core::System::GetInstance()};
     InputCommon::InputSubsystem input_subsystem;