vulkan_device: Blacklist RADV on RDNA2 from VK_EXT_vertex_input_dynamic_state

RDNA2 devices running under the RADV driver were crashing when VK_EXT_vertex_input_dynamic_state was enabled.

Blacklisting these devices until a proper fix is established.

CMakeLists.txt

@@ -33,6 +33,10 @@ option(ENABLE_CUBEB "Enables the cubeb audio backend" ON)
 
 option(USE_DISCORD_PRESENCE "Enables Discord Rich Presence" OFF)
 
+option(YUZU_USE_BUNDLED_OPUS "Compile bundled opus" ON)
+
+option(YUZU_TESTS "Compile tests" ON)
+
 # Default to a Release build
 get_property(IS_MULTI_CONFIG GLOBAL PROPERTY GENERATOR_IS_MULTI_CONFIG)
 if (NOT IS_MULTI_CONFIG AND NOT CMAKE_BUILD_TYPE)
@@ -166,8 +170,7 @@ macro(yuzu_find_packages)
     # Capitalization matters here. We need the naming to match the generated paths from Conan
     set(REQUIRED_LIBS
     #    Cmake Pkg Prefix  Version     Conan Pkg
-        "Catch2            2.13.7      catch2/2.13.7"
-        "fmt               8.0.1       fmt/8.0.1"
+        "fmt               8.0.1       fmt/8.1.1"
         "lz4               1.8         lz4/1.9.2"
         "nlohmann_json     3.8         nlohmann_json/3.8.0"
         "ZLIB              1.2         zlib/1.2.11"
@@ -175,6 +178,11 @@ macro(yuzu_find_packages)
     # can't use opus until AVX check is fixed: https://github.com/yuzu-emu/yuzu/pull/4068
         #"opus              1.3         opus/1.3.1"
     )
+    if (YUZU_TESTS)
+        list(APPEND REQUIRED_LIBS
+        "Catch2            2.13.7      catch2/2.13.7"
+        )
+    endif()
 
     foreach(PACKAGE ${REQUIRED_LIBS})
         string(REGEX REPLACE "[ \t\r\n]+" ";" PACKAGE_SPLIT ${PACKAGE})

externals/CMakeLists.txt

@@ -13,10 +13,6 @@ if (ARCHITECTURE_x86 OR ARCHITECTURE_x86_64)
     target_compile_definitions(xbyak INTERFACE XBYAK_NO_OP_NAMES)
 endif()
 
-# Catch
-add_library(catch-single-include INTERFACE)
-target_include_directories(catch-single-include INTERFACE catch/single_include)
-
 # Dynarmic
 if (ARCHITECTURE_x86_64)
     set(DYNARMIC_TESTS OFF)
@@ -117,10 +113,10 @@ if (ENABLE_WEB_SERVICE)
 endif()
 
 # Opus
-find_package(opus 1.3)
-if (NOT opus_FOUND)
-    message(STATUS "opus 1.3 or newer not found, falling back to externals")
+if (YUZU_USE_BUNDLED_OPUS)
     add_subdirectory(opus EXCLUDE_FROM_ALL)
+else()
+    find_package(opus 1.3 REQUIRED)
 endif()
 
 # FFMpeg

externals/FidelityFX-FSR/ffx-fsr/ffx_fsr1.h

@@ -747,12 +747,12 @@ AF1 sharpness){
   // Immediate constants for peak range.
   AF2 peakC=AF2(1.0,-1.0*4.0);
   // Limiters, these need to be high precision RCPs.
-  AF1 hitMinR=mn4R*ARcpF1(AF1_(4.0)*mx4R);
-  AF1 hitMinG=mn4G*ARcpF1(AF1_(4.0)*mx4G);
-  AF1 hitMinB=mn4B*ARcpF1(AF1_(4.0)*mx4B);
-  AF1 hitMaxR=(peakC.x-mx4R)*ARcpF1(AF1_(4.0)*mn4R+peakC.y);
-  AF1 hitMaxG=(peakC.x-mx4G)*ARcpF1(AF1_(4.0)*mn4G+peakC.y);
-  AF1 hitMaxB=(peakC.x-mx4B)*ARcpF1(AF1_(4.0)*mn4B+peakC.y);
+  AF1 hitMinR=min(mn4R,eR)*ARcpF1(AF1_(4.0)*mx4R);
+  AF1 hitMinG=min(mn4G,eG)*ARcpF1(AF1_(4.0)*mx4G);
+  AF1 hitMinB=min(mn4B,eB)*ARcpF1(AF1_(4.0)*mx4B);
+  AF1 hitMaxR=(peakC.x-max(mx4R,eR))*ARcpF1(AF1_(4.0)*mn4R+peakC.y);
+  AF1 hitMaxG=(peakC.x-max(mx4G,eG))*ARcpF1(AF1_(4.0)*mn4G+peakC.y);
+  AF1 hitMaxB=(peakC.x-max(mx4B,eB))*ARcpF1(AF1_(4.0)*mn4B+peakC.y);
   AF1 lobeR=max(-hitMinR,hitMaxR);
   AF1 lobeG=max(-hitMinG,hitMaxG);
   AF1 lobeB=max(-hitMinB,hitMaxB);
@@ -845,12 +845,12 @@ AF1 sharpness){
   // Immediate constants for peak range.
   AH2 peakC=AH2(1.0,-1.0*4.0);
   // Limiters, these need to be high precision RCPs.
-  AH1 hitMinR=mn4R*ARcpH1(AH1_(4.0)*mx4R);
-  AH1 hitMinG=mn4G*ARcpH1(AH1_(4.0)*mx4G);
-  AH1 hitMinB=mn4B*ARcpH1(AH1_(4.0)*mx4B);
-  AH1 hitMaxR=(peakC.x-mx4R)*ARcpH1(AH1_(4.0)*mn4R+peakC.y);
-  AH1 hitMaxG=(peakC.x-mx4G)*ARcpH1(AH1_(4.0)*mn4G+peakC.y);
-  AH1 hitMaxB=(peakC.x-mx4B)*ARcpH1(AH1_(4.0)*mn4B+peakC.y);
+  AH1 hitMinR=min(mn4R,eR)*ARcpH1(AH1_(4.0)*mx4R);
+  AH1 hitMinG=min(mn4G,eG)*ARcpH1(AH1_(4.0)*mx4G);
+  AH1 hitMinB=min(mn4B,eB)*ARcpH1(AH1_(4.0)*mx4B);
+  AH1 hitMaxR=(peakC.x-max(mx4R,eR))*ARcpH1(AH1_(4.0)*mn4R+peakC.y);
+  AH1 hitMaxG=(peakC.x-max(mx4G,eG))*ARcpH1(AH1_(4.0)*mn4G+peakC.y);
+  AH1 hitMaxB=(peakC.x-max(mx4B,eB))*ARcpH1(AH1_(4.0)*mn4B+peakC.y);
   AH1 lobeR=max(-hitMinR,hitMaxR);
   AH1 lobeG=max(-hitMinG,hitMaxG);
   AH1 lobeB=max(-hitMinB,hitMaxB);
@@ -963,12 +963,12 @@ AF1 sharpness){
   // Immediate constants for peak range.
   AH2 peakC=AH2(1.0,-1.0*4.0);
   // Limiters, these need to be high precision RCPs.
-  AH2 hitMinR=mn4R*ARcpH2(AH2_(4.0)*mx4R);
-  AH2 hitMinG=mn4G*ARcpH2(AH2_(4.0)*mx4G);
-  AH2 hitMinB=mn4B*ARcpH2(AH2_(4.0)*mx4B);
-  AH2 hitMaxR=(peakC.x-mx4R)*ARcpH2(AH2_(4.0)*mn4R+peakC.y);
-  AH2 hitMaxG=(peakC.x-mx4G)*ARcpH2(AH2_(4.0)*mn4G+peakC.y);
-  AH2 hitMaxB=(peakC.x-mx4B)*ARcpH2(AH2_(4.0)*mn4B+peakC.y);
+  AH2 hitMinR=min(mn4R,eR)*ARcpH2(AH2_(4.0)*mx4R);
+  AH2 hitMinG=min(mn4G,eG)*ARcpH2(AH2_(4.0)*mx4G);
+  AH2 hitMinB=min(mn4B,eB)*ARcpH2(AH2_(4.0)*mx4B);
+  AH2 hitMaxR=(peakC.x-max(mx4R,eR))*ARcpH2(AH2_(4.0)*mn4R+peakC.y);
+  AH2 hitMaxG=(peakC.x-max(mx4G,eG))*ARcpH2(AH2_(4.0)*mn4G+peakC.y);
+  AH2 hitMaxB=(peakC.x-max(mx4B,eB))*ARcpH2(AH2_(4.0)*mn4B+peakC.y);
   AH2 lobeR=max(-hitMinR,hitMaxR);
   AH2 lobeG=max(-hitMinG,hitMaxG);
   AH2 lobeB=max(-hitMinB,hitMaxB);

externals/ffmpeg/CMakeLists.txt

@@ -95,7 +95,8 @@ if (NOT WIN32)
         # ffnvenc could load CUDA libraries at the runtime using dlopen/dlsym or LoadLibrary/GetProcAddress
         # here we handle the hard-linking senario where CUDA is linked during compilation
         if (CUDA_FOUND)
-            list(APPEND FFmpeg_HWACCEL_FLAGS --extra-cflags=-I${CUDA_INCLUDE_DIRS})
+            # This line causes build error if CUDA_INCLUDE_DIRS is anything but a single non-empty value
+            #list(APPEND FFmpeg_HWACCEL_FLAGS --extra-cflags=-I${CUDA_INCLUDE_DIRS})
             list(APPEND FFmpeg_HWACCEL_LIBRARIES ${CUDA_LIBRARIES})
             list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${CUDA_INCLUDE_DIRS})
             list(APPEND FFmpeg_HWACCEL_LDFLAGS ${CUDA_LDFLAGS})
@@ -119,6 +120,8 @@ if (NOT WIN32)
 
     # `configure` parameters builds only exactly what yuzu needs from FFmpeg
     # `--disable-vdpau` is needed to avoid linking issues
+    set(FFmpeg_CC ${CMAKE_C_COMPILER_LAUNCHER}  ${CMAKE_C_COMPILER})
+    set(FFmpeg_CXX ${CMAKE_CXX_COMPILER_LAUNCHER}  ${CMAKE_CXX_COMPILER})
     add_custom_command(
         OUTPUT
             ${FFmpeg_MAKEFILE}
@@ -137,12 +140,14 @@ if (NOT WIN32)
                 --enable-decoder=h264
                 --enable-decoder=vp8
                 --enable-decoder=vp9
-                --cc="${CMAKE_C_COMPILER}"
-                --cxx="${CMAKE_CXX_COMPILER}"
+                --cc="${FFmpeg_CC}"
+                --cxx="${FFmpeg_CXX}"
                 ${FFmpeg_HWACCEL_FLAGS}
         WORKING_DIRECTORY
             ${FFmpeg_BUILD_DIR}
     )
+    unset(FFmpeg_CC)
+    unset(FFmpeg_CXX)
     unset(FFmpeg_HWACCEL_FLAGS)
 
     # Workaround for Ubuntu 18.04's older version of make not being able to call make as a child

src/CMakeLists.txt

@@ -151,7 +151,10 @@ add_subdirectory(audio_core)
 add_subdirectory(video_core)
 add_subdirectory(input_common)
 add_subdirectory(shader_recompiler)
-add_subdirectory(tests)
+
+if (YUZU_TESTS)
+    add_subdirectory(tests)
+endif()
 
 if (ENABLE_SDL2)
     add_subdirectory(yuzu_cmd)

src/audio_core/common.h

@@ -15,7 +15,9 @@ constexpr ResultCode ERR_INVALID_PARAMETERS{ErrorModule::Audio, 41};
 constexpr ResultCode ERR_SPLITTER_SORT_FAILED{ErrorModule::Audio, 43};
 } // namespace Audren
 
-constexpr u32_le CURRENT_PROCESS_REVISION = Common::MakeMagic('R', 'E', 'V', '9');
+constexpr u8 BASE_REVISION = '0';
+constexpr u32_le CURRENT_PROCESS_REVISION =
+    Common::MakeMagic('R', 'E', 'V', static_cast<u8>(BASE_REVISION + 0xA));
 constexpr std::size_t MAX_MIX_BUFFERS = 24;
 constexpr std::size_t MAX_BIQUAD_FILTERS = 2;
 constexpr std::size_t MAX_CHANNEL_COUNT = 6;

src/audio_core/stream.cpp

@@ -79,8 +79,8 @@ static void VolumeAdjustSamples(std::vector<s16>& samples, float game_volume) {
         return;
     }
 
-    // Implementation of a volume slider with a dynamic range of 60 dB
-    const float volume_scale_factor = volume == 0 ? 0 : std::exp(6.90775f * volume) * 0.001f;
+    // Perceived volume is not the same as the volume level
+    const float volume_scale_factor = (0.85f * ((volume * volume) - volume)) + volume;
     for (auto& sample : samples) {
         sample = static_cast<s16>(sample * volume_scale_factor);
     }

src/common/CMakeLists.txt

@@ -85,6 +85,7 @@ add_library(common STATIC
     logging/backend.h
     logging/filter.cpp
     logging/filter.h
+    logging/formatter.h
     logging/log.h
     logging/log_entry.h
     logging/text_formatter.cpp

src/common/bit_util.h

@@ -45,6 +45,12 @@ template <typename T>
     return static_cast<u32>(log2_f + static_cast<u64>((value ^ (1ULL << log2_f)) != 0ULL));
 }
 
+template <typename T>
+requires std::is_unsigned_v<T>
+[[nodiscard]] constexpr bool IsPow2(T value) {
+    return std::has_single_bit(value);
+}
+
 template <typename T>
 requires std::is_integral_v<T>
 [[nodiscard]] T NextPow2(T value) {

src/common/common_types.h

@@ -46,13 +46,3 @@ using GPUVAddr = u64; ///< Represents a pointer in the GPU virtual address space
 
 using u128 = std::array<std::uint64_t, 2>;
 static_assert(sizeof(u128) == 16, "u128 must be 128 bits wide");
-
-// An inheritable class to disallow the copy constructor and operator= functions
-class NonCopyable {
-protected:
-    constexpr NonCopyable() = default;
-    ~NonCopyable() = default;
-
-    NonCopyable(const NonCopyable&) = delete;
-    NonCopyable& operator=(const NonCopyable&) = delete;
-};

src/common/fiber.cpp

@@ -124,7 +124,10 @@ void Fiber::YieldTo(std::weak_ptr<Fiber> weak_from, Fiber& to) {
 
     // "from" might no longer be valid if the thread was killed
     if (auto from = weak_from.lock()) {
-        ASSERT(from->impl->previous_fiber != nullptr);
+        if (from->impl->previous_fiber == nullptr) {
+            ASSERT_MSG(false, "previous_fiber is nullptr!");
+            return;
+        }
         from->impl->previous_fiber->impl->context = transfer.fctx;
         from->impl->previous_fiber->impl->guard.unlock();
         from->impl->previous_fiber.reset();

src/common/fs/file.h

@@ -188,9 +188,8 @@ public:
 
 #ifdef _WIN32
     template <typename Path>
-    [[nodiscard]] void Open(const Path& path, FileAccessMode mode,
-                            FileType type = FileType::BinaryFile,
-                            FileShareFlag flag = FileShareFlag::ShareReadOnly) {
+    void Open(const Path& path, FileAccessMode mode, FileType type = FileType::BinaryFile,
+              FileShareFlag flag = FileShareFlag::ShareReadOnly) {
         using ValueType = typename Path::value_type;
         if constexpr (IsChar<ValueType>) {
             Open(ToU8String(path), mode, type, flag);

src/common/fs/fs_util.cpp

@@ -16,6 +16,10 @@ std::u8string BufferToU8String(std::span<const u8> buffer) {
     return std::u8string{buffer.begin(), std::ranges::find(buffer, u8{0})};
 }
 
+std::u8string_view BufferToU8StringView(std::span<const u8> buffer) {
+    return std::u8string_view{reinterpret_cast<const char8_t*>(buffer.data())};
+}
+
 std::string ToUTF8String(std::u8string_view u8_string) {
     return std::string{u8_string.begin(), u8_string.end()};
 }
@@ -24,6 +28,10 @@ std::string BufferToUTF8String(std::span<const u8> buffer) {
     return std::string{buffer.begin(), std::ranges::find(buffer, u8{0})};
 }
 
+std::string_view BufferToUTF8StringView(std::span<const u8> buffer) {
+    return std::string_view{reinterpret_cast<const char*>(buffer.data())};
+}
+
 std::string PathToUTF8String(const std::filesystem::path& path) {
     return ToUTF8String(path.u8string());
 }

src/common/fs/fs_util.h

@@ -37,6 +37,15 @@ concept IsChar = std::same_as<T, char>;
  */
 [[nodiscard]] std::u8string BufferToU8String(std::span<const u8> buffer);
 
+/**
+ * Same as BufferToU8String, but returns a string view of the buffer.
+ *
+ * @param buffer Buffer of bytes
+ *
+ * @returns UTF-8 encoded std::u8string_view.
+ */
+[[nodiscard]] std::u8string_view BufferToU8StringView(std::span<const u8> buffer);
+
 /**
  * Converts a std::u8string or std::u8string_view to a UTF-8 encoded std::string.
  *
@@ -57,6 +66,15 @@ concept IsChar = std::same_as<T, char>;
  */
 [[nodiscard]] std::string BufferToUTF8String(std::span<const u8> buffer);
 
+/**
+ * Same as BufferToUTF8String, but returns a string view of the buffer.
+ *
+ * @param buffer Buffer of bytes
+ *
+ * @returns UTF-8 encoded std::string_view.
+ */
+[[nodiscard]] std::string_view BufferToUTF8StringView(std::span<const u8> buffer);
+
 /**
  * Converts a filesystem path to a UTF-8 encoded std::string.
  *

src/common/input.h

@@ -209,6 +209,15 @@ enum class ButtonNames {
     Triangle,
     Share,
     Options,
+    Home,
+    Touch,
+
+    // Mouse buttons
+    ButtonMouseWheel,
+    ButtonBackward,
+    ButtonForward,
+    ButtonTask,
+    ButtonExtra,
 };
 
 // Callback data consisting of an input type and the equivalent data status

src/common/logging/filter.cpp

@@ -108,6 +108,7 @@ bool ParseFilterRule(Filter& instance, Iterator begin, Iterator end) {
     SUB(Service, Migration)                                                                        \
     SUB(Service, Mii)                                                                              \
     SUB(Service, MM)                                                                               \
+    SUB(Service, MNPP)                                                                             \
     SUB(Service, NCM)                                                                              \
     SUB(Service, NFC)                                                                              \
     SUB(Service, NFP)                                                                              \

src/common/logging/formatter.h

@@ -0,0 +1,23 @@
+// Copyright 2022 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <type_traits>
+
+#include <fmt/format.h>
+
+// adapted from https://github.com/fmtlib/fmt/issues/2704
+// a generic formatter for enum classes
+#if FMT_VERSION >= 80100
+template <typename T>
+struct fmt::formatter<T, std::enable_if_t<std::is_enum_v<T>, char>>
+    : formatter<std::underlying_type_t<T>> {
+    template <typename FormatContext>
+    auto format(const T& value, FormatContext& ctx) -> decltype(ctx.out()) {
+        return fmt::formatter<std::underlying_type_t<T>>::format(
+            static_cast<std::underlying_type_t<T>>(value), ctx);
+    }
+};
+#endif

src/common/logging/log.h

@@ -7,8 +7,9 @@
 #include <algorithm>
 #include <string_view>
 
-#include <fmt/core.h>
+#include <fmt/format.h>
 
+#include "common/logging/formatter.h"
 #include "common/logging/types.h"
 
 namespace Common::Log {

src/common/logging/types.h

@@ -76,6 +76,7 @@ enum class Class : u8 {
     Service_Migration, ///< The migration service
     Service_Mii,       ///< The Mii service
     Service_MM,        ///< The MM (Multimedia) service
+    Service_MNPP,      ///< The MNPP service
     Service_NCM,       ///< The NCM service
     Service_NFC,       ///< The NFC (Near-field communication) service
     Service_NFP,       ///< The NFP service

src/common/page_table.cpp

@@ -10,11 +10,65 @@ PageTable::PageTable() = default;
 
 PageTable::~PageTable() noexcept = default;
 
-void PageTable::Resize(size_t address_space_width_in_bits, size_t page_size_in_bits) {
-    const size_t num_page_table_entries{1ULL << (address_space_width_in_bits - page_size_in_bits)};
+bool PageTable::BeginTraversal(TraversalEntry& out_entry, TraversalContext& out_context,
+                               u64 address) const {
+    // Setup invalid defaults.
+    out_entry.phys_addr = 0;
+    out_entry.block_size = page_size;
+    out_context.next_page = 0;
+
+    // Validate that we can read the actual entry.
+    const auto page = address / page_size;
+    if (page >= backing_addr.size()) {
+        return false;
+    }
+
+    // Validate that the entry is mapped.
+    const auto phys_addr = backing_addr[page];
+    if (phys_addr == 0) {
+        return false;
+    }
+
+    // Populate the results.
+    out_entry.phys_addr = phys_addr + address;
+    out_context.next_page = page + 1;
+    out_context.next_offset = address + page_size;
+
+    return true;
+}
+
+bool PageTable::ContinueTraversal(TraversalEntry& out_entry, TraversalContext& context) const {
+    // Setup invalid defaults.
+    out_entry.phys_addr = 0;
+    out_entry.block_size = page_size;
+
+    // Validate that we can read the actual entry.
+    const auto page = context.next_page;
+    if (page >= backing_addr.size()) {
+        return false;
+    }
+
+    // Validate that the entry is mapped.
+    const auto phys_addr = backing_addr[page];
+    if (phys_addr == 0) {
+        return false;
+    }
+
+    // Populate the results.
+    out_entry.phys_addr = phys_addr + context.next_offset;
+    context.next_page = page + 1;
+    context.next_offset += page_size;
+
+    return true;
+}
+
+void PageTable::Resize(std::size_t address_space_width_in_bits, std::size_t page_size_in_bits) {
+    const std::size_t num_page_table_entries{1ULL
+                                             << (address_space_width_in_bits - page_size_in_bits)};
     pointers.resize(num_page_table_entries);
     backing_addr.resize(num_page_table_entries);
     current_address_space_width_in_bits = address_space_width_in_bits;
+    page_size = 1ULL << page_size_in_bits;
 }
 
 } // namespace Common

src/common/page_table.h

@@ -27,6 +27,16 @@ enum class PageType : u8 {
  * mimics the way a real CPU page table works.
  */
 struct PageTable {
+    struct TraversalEntry {
+        u64 phys_addr{};
+        std::size_t block_size{};
+    };
+
+    struct TraversalContext {
+        u64 next_page{};
+        u64 next_offset{};
+    };
+
     /// Number of bits reserved for attribute tagging.
     /// This can be at most the guaranteed alignment of the pointers in the page table.
     static constexpr int ATTRIBUTE_BITS = 2;
@@ -89,6 +99,10 @@ struct PageTable {
     PageTable(PageTable&&) noexcept = default;
     PageTable& operator=(PageTable&&) noexcept = default;
 
+    bool BeginTraversal(TraversalEntry& out_entry, TraversalContext& out_context,
+                        u64 address) const;
+    bool ContinueTraversal(TraversalEntry& out_entry, TraversalContext& context) const;
+
     /**
      * Resizes the page table to be able to accommodate enough pages within
      * a given address space.
@@ -96,9 +110,9 @@ struct PageTable {
      * @param address_space_width_in_bits The address size width in bits.
      * @param page_size_in_bits           The page size in bits.
      */
-    void Resize(size_t address_space_width_in_bits, size_t page_size_in_bits);
+    void Resize(std::size_t address_space_width_in_bits, std::size_t page_size_in_bits);
 
-    size_t GetAddressSpaceBits() const {
+    std::size_t GetAddressSpaceBits() const {
         return current_address_space_width_in_bits;
     }
 
@@ -110,9 +124,11 @@ struct PageTable {
 
     VirtualBuffer<u64> backing_addr;
 
-    size_t current_address_space_width_in_bits;
+    std::size_t current_address_space_width_in_bits{};
 
-    u8* fastmem_arena;
+    u8* fastmem_arena{};
+
+    std::size_t page_size{};
 };
 
 } // namespace Common

src/common/settings.h

@@ -554,6 +554,7 @@ struct Values {
     Setting<bool> use_docked_mode{true, "use_docked_mode"};
 
     BasicSetting<bool> enable_raw_input{false, "enable_raw_input"};
+    BasicSetting<bool> controller_navigation{true, "controller_navigation"};
 
     Setting<bool> vibration_enabled{true, "vibration_enabled"};
     Setting<bool> enable_accurate_vibrations{false, "enable_accurate_vibrations"};
@@ -597,6 +598,7 @@ struct Values {
     BasicSetting<std::string> program_args{std::string(), "program_args"};
     BasicSetting<bool> dump_exefs{false, "dump_exefs"};
     BasicSetting<bool> dump_nso{false, "dump_nso"};
+    BasicSetting<bool> dump_shaders{false, "dump_shaders"};
     BasicSetting<bool> enable_fs_access_log{false, "enable_fs_access_log"};
     BasicSetting<bool> reporting_services{false, "reporting_services"};
     BasicSetting<bool> quest_flag{false, "quest_flag"};

src/common/telemetry.h

@@ -8,6 +8,7 @@
 #include <map>
 #include <memory>
 #include <string>
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 
 namespace Common::Telemetry {
@@ -28,7 +29,7 @@ struct VisitorInterface;
 /**
  * Interface class for telemetry data fields.
  */
-class FieldInterface : NonCopyable {
+class FieldInterface {
 public:
     virtual ~FieldInterface() = default;
 
@@ -52,14 +53,15 @@ public:
 template <typename T>
 class Field : public FieldInterface {
 public:
+    YUZU_NON_COPYABLE(Field);
+
     Field(FieldType type_, std::string name_, T value_)
         : name(std::move(name_)), type(type_), value(std::move(value_)) {}
 
-    Field(const Field&) = default;
-    Field& operator=(const Field&) = default;
+    ~Field() override = default;
 
-    Field(Field&&) = default;
-    Field& operator=(Field&& other) = default;
+    Field(Field&&) noexcept = default;
+    Field& operator=(Field&& other) noexcept = default;
 
     void Accept(VisitorInterface& visitor) const override;
 
@@ -98,9 +100,15 @@ private:
 /**
  * Collection of data fields that have been logged.
  */
-class FieldCollection final : NonCopyable {
+class FieldCollection final {
 public:
+    YUZU_NON_COPYABLE(FieldCollection);
+
     FieldCollection() = default;
+    ~FieldCollection() = default;
+
+    FieldCollection(FieldCollection&&) noexcept = default;
+    FieldCollection& operator=(FieldCollection&&) noexcept = default;
 
     /**
      * Accept method for the visitor pattern, visits each field in the collection.
@@ -133,7 +141,7 @@ private:
  * Telemetry fields visitor interface class. A backend to log to a web service should implement
  * this interface.
  */
-struct VisitorInterface : NonCopyable {
+struct VisitorInterface {
     virtual ~VisitorInterface() = default;
 
     virtual void Visit(const Field<bool>& field) = 0;
@@ -160,8 +168,11 @@ struct VisitorInterface : NonCopyable {
  * Empty implementation of VisitorInterface that drops all fields. Used when a functional
  * backend implementation is not available.
  */
-struct NullVisitor : public VisitorInterface {
-    ~NullVisitor() = default;
+struct NullVisitor final : public VisitorInterface {
+    YUZU_NON_COPYABLE(NullVisitor);
+
+    NullVisitor() = default;
+    ~NullVisitor() override = default;
 
     void Visit(const Field<bool>& /*field*/) override {}
     void Visit(const Field<double>& /*field*/) override {}

src/common/uuid.cpp

@@ -1,23 +1,25 @@
-// Copyright 2018 yuzu Emulator Project
+// Copyright 2022 yuzu Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <bit>
+#include <optional>
 #include <random>
 
 #include <fmt/format.h>
 
 #include "common/assert.h"
+#include "common/tiny_mt.h"
 #include "common/uuid.h"
 
 namespace Common {
 
 namespace {
 
-bool IsHexDigit(char c) {
-    return (c >= '0' && c <= '9') || (c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F');
-}
+constexpr size_t RawStringSize = sizeof(UUID) * 2;
+constexpr size_t FormattedStringSize = RawStringSize + 4;
 
-u8 HexCharToByte(char c) {
+std::optional<u8> HexCharToByte(char c) {
     if (c >= '0' && c <= '9') {
         return static_cast<u8>(c - '0');
     }
@@ -28,60 +30,184 @@ u8 HexCharToByte(char c) {
         return static_cast<u8>(c - 'A' + 10);
     }
     ASSERT_MSG(false, "{} is not a hexadecimal digit!", c);
-    return u8{0};
+    return std::nullopt;
+}
+
+std::array<u8, 0x10> ConstructFromRawString(std::string_view raw_string) {
+    std::array<u8, 0x10> uuid;
+
+    for (size_t i = 0; i < RawStringSize; i += 2) {
+        const auto upper = HexCharToByte(raw_string[i]);
+        const auto lower = HexCharToByte(raw_string[i + 1]);
+        if (!upper || !lower) {
+            return {};
+        }
+        uuid[i / 2] = static_cast<u8>((*upper << 4) | *lower);
+    }
+
+    return uuid;
+}
+
+std::array<u8, 0x10> ConstructFromFormattedString(std::string_view formatted_string) {
+    std::array<u8, 0x10> uuid;
+
+    size_t i = 0;
+
+    // Process the first 8 characters.
+    const auto* str = formatted_string.data();
+
+    for (; i < 4; ++i) {
+        const auto upper = HexCharToByte(*(str++));
+        const auto lower = HexCharToByte(*(str++));
+        if (!upper || !lower) {
+            return {};
+        }
+        uuid[i] = static_cast<u8>((*upper << 4) | *lower);
+    }
+
+    // Process the next 4 characters.
+    ++str;
+
+    for (; i < 6; ++i) {
+        const auto upper = HexCharToByte(*(str++));
+        const auto lower = HexCharToByte(*(str++));
+        if (!upper || !lower) {
+            return {};
+        }
+        uuid[i] = static_cast<u8>((*upper << 4) | *lower);
+    }
+
+    // Process the next 4 characters.
+    ++str;
+
+    for (; i < 8; ++i) {
+        const auto upper = HexCharToByte(*(str++));
+        const auto lower = HexCharToByte(*(str++));
+        if (!upper || !lower) {
+            return {};
+        }
+        uuid[i] = static_cast<u8>((*upper << 4) | *lower);
+    }
+
+    // Process the next 4 characters.
+    ++str;
+
+    for (; i < 10; ++i) {
+        const auto upper = HexCharToByte(*(str++));
+        const auto lower = HexCharToByte(*(str++));
+        if (!upper || !lower) {
+            return {};
+        }
+        uuid[i] = static_cast<u8>((*upper << 4) | *lower);
+    }
+
+    // Process the last 12 characters.
+    ++str;
+
+    for (; i < 16; ++i) {
+        const auto upper = HexCharToByte(*(str++));
+        const auto lower = HexCharToByte(*(str++));
+        if (!upper || !lower) {
+            return {};
+        }
+        uuid[i] = static_cast<u8>((*upper << 4) | *lower);
+    }
+
+    return uuid;
+}
+
+std::array<u8, 0x10> ConstructUUID(std::string_view uuid_string) {
+    const auto length = uuid_string.length();
+
+    if (length == 0) {
+        return {};
+    }
+
+    // Check if the input string contains 32 hexadecimal characters.
+    if (length == RawStringSize) {
+        return ConstructFromRawString(uuid_string);
+    }
+
+    // Check if the input string has the length of a RFC 4122 formatted UUID string.
+    if (length == FormattedStringSize) {
+        return ConstructFromFormattedString(uuid_string);
+    }
+
+    ASSERT_MSG(false, "UUID string has an invalid length of {} characters!", length);
+
+    return {};
 }
 
 } // Anonymous namespace
 
-u128 HexStringToU128(std::string_view hex_string) {
-    const size_t length = hex_string.length();
+UUID::UUID(std::string_view uuid_string) : uuid{ConstructUUID(uuid_string)} {}
 
-    // Detect "0x" prefix.
-    const bool has_0x_prefix = length > 2 && hex_string[0] == '0' && hex_string[1] == 'x';
-    const size_t offset = has_0x_prefix ? 2 : 0;
-
-    // Check length.
-    if (length > 32 + offset) {
-        ASSERT_MSG(false, "hex_string has more than 32 hexadecimal characters!");
-        return INVALID_UUID;
-    }
-
-    u64 lo = 0;
-    u64 hi = 0;
-    for (size_t i = 0; i < length - offset; ++i) {
-        const char c = hex_string[length - 1 - i];
-        if (!IsHexDigit(c)) {
-            ASSERT_MSG(false, "{} is not a hexadecimal digit!", c);
-            return INVALID_UUID;
-        }
-        if (i < 16) {
-            lo |= u64{HexCharToByte(c)} << (i * 4);
-        }
-        if (i >= 16) {
-            hi |= u64{HexCharToByte(c)} << ((i - 16) * 4);
-        }
-    }
-    return u128{lo, hi};
+std::string UUID::RawString() const {
+    return fmt::format("{:02x}{:02x}{:02x}{:02x}{:02x}{:02x}{:02x}{:02x}"
+                       "{:02x}{:02x}{:02x}{:02x}{:02x}{:02x}{:02x}{:02x}",
+                       uuid[0], uuid[1], uuid[2], uuid[3], uuid[4], uuid[5], uuid[6], uuid[7],
+                       uuid[8], uuid[9], uuid[10], uuid[11], uuid[12], uuid[13], uuid[14],
+                       uuid[15]);
 }
 
-UUID UUID::Generate() {
+std::string UUID::FormattedString() const {
+    return fmt::format("{:02x}{:02x}{:02x}{:02x}"
+                       "-{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-"
+                       "{:02x}{:02x}{:02x}{:02x}{:02x}{:02x}",
+                       uuid[0], uuid[1], uuid[2], uuid[3], uuid[4], uuid[5], uuid[6], uuid[7],
+                       uuid[8], uuid[9], uuid[10], uuid[11], uuid[12], uuid[13], uuid[14],
+                       uuid[15]);
+}
+
+size_t UUID::Hash() const noexcept {
+    u64 upper_hash;
+    u64 lower_hash;
+
+    std::memcpy(&upper_hash, uuid.data(), sizeof(u64));
+    std::memcpy(&lower_hash, uuid.data() + sizeof(u64), sizeof(u64));
+
+    return upper_hash ^ std::rotl(lower_hash, 1);
+}
+
+u128 UUID::AsU128() const {
+    u128 uuid_old;
+    std::memcpy(&uuid_old, uuid.data(), sizeof(UUID));
+    return uuid_old;
+}
+
+UUID UUID::MakeRandom() {
     std::random_device device;
-    std::mt19937 gen(device());
-    std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max());
-    return UUID{distribution(gen), distribution(gen)};
+
+    return MakeRandomWithSeed(device());
 }
 
-std::string UUID::Format() const {
-    return fmt::format("{:016x}{:016x}", uuid[1], uuid[0]);
+UUID UUID::MakeRandomWithSeed(u32 seed) {
+    // Create and initialize our RNG.
+    TinyMT rng;
+    rng.Initialize(seed);
+
+    UUID uuid;
+
+    // Populate the UUID with random bytes.
+    rng.GenerateRandomBytes(uuid.uuid.data(), sizeof(UUID));
+
+    return uuid;
 }
 
-std::string UUID::FormatSwitch() const {
-    std::array<u8, 16> s{};
-    std::memcpy(s.data(), uuid.data(), sizeof(u128));
-    return fmt::format("{:02x}{:02x}{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{"
-                       ":02x}{:02x}{:02x}{:02x}{:02x}",
-                       s[0], s[1], s[2], s[3], s[4], s[5], s[6], s[7], s[8], s[9], s[10], s[11],
-                       s[12], s[13], s[14], s[15]);
+UUID UUID::MakeRandomRFC4122V4() {
+    auto uuid = MakeRandom();
+
+    // According to Proposed Standard RFC 4122 Section 4.4, we must:
+
+    // 1. Set the two most significant bits (bits 6 and 7) of the
+    //    clock_seq_hi_and_reserved to zero and one, respectively.
+    uuid.uuid[8] = 0x80 | (uuid.uuid[8] & 0x3F);
+
+    // 2. Set the four most significant bits (bits 12 through 15) of the
+    //    time_hi_and_version field to the 4-bit version number from Section 4.1.3.
+    uuid.uuid[6] = 0x40 | (uuid.uuid[6] & 0xF);
+
+    return uuid;
 }
 
 } // namespace Common

src/common/uuid.h

@@ -1,9 +1,11 @@
-// Copyright 2018 yuzu Emulator Project
+// Copyright 2022 yuzu Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
 #pragma once
 
+#include <array>
+#include <functional>
 #include <string>
 #include <string_view>
 
@@ -11,69 +13,119 @@
 
 namespace Common {
 
-constexpr u128 INVALID_UUID{{0, 0}};
-
-/**
- * Converts a hex string to a 128-bit unsigned integer.
- *
- * The hex string can be formatted in lowercase or uppercase, with or without the "0x" prefix.
- *
- * This function will assert and return INVALID_UUID under the following conditions:
- * - If the hex string is more than 32 characters long
- * - If the hex string contains non-hexadecimal characters
- *
- * @param hex_string Hexadecimal string
- *
- * @returns A 128-bit unsigned integer if successfully converted, INVALID_UUID otherwise.
- */
-[[nodiscard]] u128 HexStringToU128(std::string_view hex_string);
-
 struct UUID {
-    // UUIDs which are 0 are considered invalid!
-    u128 uuid;
-    UUID() = default;
-    constexpr explicit UUID(const u128& id) : uuid{id} {}
-    constexpr explicit UUID(const u64 lo, const u64 hi) : uuid{{lo, hi}} {}
-    explicit UUID(std::string_view hex_string) {
-        uuid = HexStringToU128(hex_string);
+    std::array<u8, 0x10> uuid{};
+
+    /// Constructs an invalid UUID.
+    constexpr UUID() = default;
+
+    /// Constructs a UUID from a reference to a 128 bit array.
+    constexpr explicit UUID(const std::array<u8, 16>& uuid_) : uuid{uuid_} {}
+
+    /**
+     * Constructs a UUID from either:
+     * 1. A 32 hexadecimal character string representing the bytes of the UUID
+     * 2. A RFC 4122 formatted UUID string, in the format xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
+     *
+     * The input string may contain uppercase or lowercase characters, but they must:
+     * 1. Contain valid hexadecimal characters (0-9, a-f, A-F)
+     * 2. Not contain the "0x" hexadecimal prefix
+     *
+     * Should the input string not meet the above requirements,
+     * an assert will be triggered and an invalid UUID is set instead.
+     */
+    explicit UUID(std::string_view uuid_string);
+
+    ~UUID() = default;
+
+    constexpr UUID(const UUID&) noexcept = default;
+    constexpr UUID(UUID&&) noexcept = default;
+
+    constexpr UUID& operator=(const UUID&) noexcept = default;
+    constexpr UUID& operator=(UUID&&) noexcept = default;
+
+    /**
+     * Returns whether the stored UUID is valid or not.
+     *
+     * @returns True if the stored UUID is valid, false otherwise.
+     */
+    constexpr bool IsValid() const {
+        return uuid != std::array<u8, 0x10>{};
     }
 
-    [[nodiscard]] constexpr explicit operator bool() const {
-        return uuid != INVALID_UUID;
+    /**
+     * Returns whether the stored UUID is invalid or not.
+     *
+     * @returns True if the stored UUID is invalid, false otherwise.
+     */
+    constexpr bool IsInvalid() const {
+        return !IsValid();
     }
 
-    [[nodiscard]] constexpr bool operator==(const UUID& rhs) const {
-        return uuid == rhs.uuid;
+    /**
+     * Returns a 32 hexadecimal character string representing the bytes of the UUID.
+     *
+     * @returns A 32 hexadecimal character string of the UUID.
+     */
+    std::string RawString() const;
+
+    /**
+     * Returns a RFC 4122 formatted UUID string in the format xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx.
+     *
+     * @returns A RFC 4122 formatted UUID string.
+     */
+    std::string FormattedString() const;
+
+    /**
+     * Returns a 64-bit hash of the UUID for use in hash table data structures.
+     *
+     * @returns A 64-bit hash of the UUID.
+     */
+    size_t Hash() const noexcept;
+
+    /// DO NOT USE. Copies the contents of the UUID into a u128.
+    u128 AsU128() const;
+
+    /**
+     * Creates a default UUID "yuzu Default UID".
+     *
+     * @returns A UUID with its bytes set to the ASCII values of "yuzu Default UID".
+     */
+    static constexpr UUID MakeDefault() {
+        return UUID{
+            {'y', 'u', 'z', 'u', ' ', 'D', 'e', 'f', 'a', 'u', 'l', 't', ' ', 'U', 'I', 'D'},
+        };
     }
 
-    [[nodiscard]] constexpr bool operator!=(const UUID& rhs) const {
-        return !operator==(rhs);
-    }
+    /**
+     * Creates a random UUID.
+     *
+     * @returns A random UUID.
+     */
+    static UUID MakeRandom();
 
-    // TODO(ogniK): Properly generate uuids based on RFC-4122
-    [[nodiscard]] static UUID Generate();
+    /**
+     * Creates a random UUID with a seed.
+     *
+     * @param seed A seed to initialize the Mersenne-Twister RNG
+     *
+     * @returns A random UUID.
+     */
+    static UUID MakeRandomWithSeed(u32 seed);
 
-    // Set the UUID to {0,0} to be considered an invalid user
-    constexpr void Invalidate() {
-        uuid = INVALID_UUID;
-    }
+    /**
+     * Creates a random UUID. The generated UUID is RFC 4122 Version 4 compliant.
+     *
+     * @returns A random UUID that is RFC 4122 Version 4 compliant.
+     */
+    static UUID MakeRandomRFC4122V4();
 
-    [[nodiscard]] constexpr bool IsInvalid() const {
-        return uuid == INVALID_UUID;
-    }
-    [[nodiscard]] constexpr bool IsValid() const {
-        return !IsInvalid();
-    }
-
-    // TODO(ogniK): Properly generate a Nintendo ID
-    [[nodiscard]] constexpr u64 GetNintendoID() const {
-        return uuid[0];
-    }
-
-    [[nodiscard]] std::string Format() const;
-    [[nodiscard]] std::string FormatSwitch() const;
+    friend constexpr bool operator==(const UUID& lhs, const UUID& rhs) = default;
 };
-static_assert(sizeof(UUID) == 16, "UUID is an invalid size!");
+static_assert(sizeof(UUID) == 0x10, "UUID has incorrect size.");
+
+/// An invalid UUID. This UUID has all its bytes set to 0.
+constexpr UUID InvalidUUID = {};
 
 } // namespace Common
 
@@ -82,7 +134,7 @@ namespace std {
 template <>
 struct hash<Common::UUID> {
     size_t operator()(const Common::UUID& uuid) const noexcept {
-        return uuid.uuid[1] ^ uuid.uuid[0];
+        return uuid.Hash();
     }
 };
 

src/common/wall_clock.cpp

@@ -65,14 +65,20 @@ private:
 
 #ifdef ARCHITECTURE_x86_64
 
-std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency,
-                                                   u32 emulated_clock_frequency) {
+std::unique_ptr<WallClock> CreateBestMatchingClock(u64 emulated_cpu_frequency,
+                                                   u64 emulated_clock_frequency) {
     const auto& caps = GetCPUCaps();
     u64 rtsc_frequency = 0;
     if (caps.invariant_tsc) {
         rtsc_frequency = EstimateRDTSCFrequency();
     }
-    if (rtsc_frequency == 0) {
+
+    // Fallback to StandardWallClock if the hardware TSC does not have the precision greater than:
+    // - A nanosecond
+    // - The emulated CPU frequency
+    // - The emulated clock counter frequency (CNTFRQ)
+    if (rtsc_frequency <= WallClock::NS_RATIO || rtsc_frequency <= emulated_cpu_frequency ||
+        rtsc_frequency <= emulated_clock_frequency) {
         return std::make_unique<StandardWallClock>(emulated_cpu_frequency,
                                                    emulated_clock_frequency);
     } else {
@@ -83,8 +89,8 @@ std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency,
 
 #else
 
-std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency,
-                                                   u32 emulated_clock_frequency) {
+std::unique_ptr<WallClock> CreateBestMatchingClock(u64 emulated_cpu_frequency,
+                                                   u64 emulated_clock_frequency) {
     return std::make_unique<StandardWallClock>(emulated_cpu_frequency, emulated_clock_frequency);
 }
 

src/common/wall_clock.h

@@ -13,6 +13,10 @@ namespace Common {
 
 class WallClock {
 public:
+    static constexpr u64 NS_RATIO = 1'000'000'000;
+    static constexpr u64 US_RATIO = 1'000'000;
+    static constexpr u64 MS_RATIO = 1'000;
+
     virtual ~WallClock() = default;
 
     /// Returns current wall time in nanoseconds
@@ -49,7 +53,7 @@ private:
     bool is_native;
 };
 
-[[nodiscard]] std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency,
-                                                                 u32 emulated_clock_frequency);
+[[nodiscard]] std::unique_ptr<WallClock> CreateBestMatchingClock(u64 emulated_cpu_frequency,
+                                                                 u64 emulated_clock_frequency);
 
 } // namespace Common

src/common/x64/native_clock.cpp

@@ -47,9 +47,9 @@ NativeClock::NativeClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequen
     _mm_mfence();
     time_point.inner.last_measure = __rdtsc();
     time_point.inner.accumulated_ticks = 0U;
-    ns_rtsc_factor = GetFixedPoint64Factor(1000000000, rtsc_frequency);
-    us_rtsc_factor = GetFixedPoint64Factor(1000000, rtsc_frequency);
-    ms_rtsc_factor = GetFixedPoint64Factor(1000, rtsc_frequency);
+    ns_rtsc_factor = GetFixedPoint64Factor(NS_RATIO, rtsc_frequency);
+    us_rtsc_factor = GetFixedPoint64Factor(US_RATIO, rtsc_frequency);
+    ms_rtsc_factor = GetFixedPoint64Factor(MS_RATIO, rtsc_frequency);
     clock_rtsc_factor = GetFixedPoint64Factor(emulated_clock_frequency, rtsc_frequency);
     cpu_rtsc_factor = GetFixedPoint64Factor(emulated_cpu_frequency, rtsc_frequency);
 }

src/common/x64/xbyak_abi.h

@@ -37,12 +37,12 @@ constexpr Xbyak::Reg IndexToReg(size_t reg_index) {
     }
 }
 
-inline std::bitset<32> BuildRegSet(std::initializer_list<Xbyak::Reg> regs) {
-    std::bitset<32> bits;
+constexpr std::bitset<32> BuildRegSet(std::initializer_list<Xbyak::Reg> regs) {
+    size_t bits = 0;
     for (const Xbyak::Reg& reg : regs) {
-        bits[RegToIndex(reg)] = true;
+        bits |= size_t{1} << RegToIndex(reg);
     }
-    return bits;
+    return {bits};
 }
 
 constexpr inline std::bitset<32> ABI_ALL_GPRS(0x0000FFFF);
@@ -57,7 +57,7 @@ constexpr inline Xbyak::Reg ABI_PARAM2 = Xbyak::util::rdx;
 constexpr inline Xbyak::Reg ABI_PARAM3 = Xbyak::util::r8;
 constexpr inline Xbyak::Reg ABI_PARAM4 = Xbyak::util::r9;
 
-const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
+constexpr inline std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
     // GPRs
     Xbyak::util::rcx,
     Xbyak::util::rdx,
@@ -74,7 +74,7 @@ const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
     Xbyak::util::xmm5,
 });
 
-const std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({
+constexpr inline std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({
     // GPRs
     Xbyak::util::rbx,
     Xbyak::util::rsi,
@@ -108,7 +108,7 @@ constexpr inline Xbyak::Reg ABI_PARAM2 = Xbyak::util::rsi;
 constexpr inline Xbyak::Reg ABI_PARAM3 = Xbyak::util::rdx;
 constexpr inline Xbyak::Reg ABI_PARAM4 = Xbyak::util::rcx;
 
-const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
+constexpr inline std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
     // GPRs
     Xbyak::util::rcx,
     Xbyak::util::rdx,
@@ -137,7 +137,7 @@ const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({
     Xbyak::util::xmm15,
 });
 
-const std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({
+constexpr inline std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({
     // GPRs
     Xbyak::util::rbx,
     Xbyak::util::rbp,

src/core/CMakeLists.txt

@@ -187,6 +187,8 @@ add_library(core STATIC
     hle/kernel/k_event.h
     hle/kernel/k_handle_table.cpp
     hle/kernel/k_handle_table.h
+    hle/kernel/k_interrupt_manager.cpp
+    hle/kernel/k_interrupt_manager.h
     hle/kernel/k_light_condition_variable.cpp
     hle/kernel/k_light_condition_variable.h
     hle/kernel/k_light_lock.cpp
@@ -245,6 +247,9 @@ add_library(core STATIC
     hle/kernel/k_trace.h
     hle/kernel/k_transfer_memory.cpp
     hle/kernel/k_transfer_memory.h
+    hle/kernel/k_worker_task.h
+    hle/kernel/k_worker_task_manager.cpp
+    hle/kernel/k_worker_task_manager.h
     hle/kernel/k_writable_event.cpp
     hle/kernel/k_writable_event.h
     hle/kernel/kernel.cpp
@@ -462,6 +467,8 @@ add_library(core STATIC
     hle/service/mii/types.h
     hle/service/mm/mm_u.cpp
     hle/service/mm/mm_u.h
+    hle/service/mnpp/mnpp_app.cpp
+    hle/service/mnpp/mnpp_app.h
     hle/service/ncm/ncm.cpp
     hle/service/ncm/ncm.h
     hle/service/nfc/nfc.cpp

src/core/arm/arm_interface.h

@@ -6,6 +6,7 @@
 
 #include <array>
 #include <vector>
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "core/hardware_properties.h"
 
@@ -24,8 +25,11 @@ class CPUInterruptHandler;
 using CPUInterrupts = std::array<CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES>;
 
 /// Generic ARMv8 CPU interface
-class ARM_Interface : NonCopyable {
+class ARM_Interface {
 public:
+    YUZU_NON_COPYABLE(ARM_Interface);
+    YUZU_NON_MOVEABLE(ARM_Interface);
+
     explicit ARM_Interface(System& system_, CPUInterrupts& interrupt_handlers_,
                            bool uses_wall_clock_)
         : system{system_}, interrupt_handlers{interrupt_handlers_}, uses_wall_clock{

src/core/core.cpp

@@ -317,6 +317,8 @@ struct System::Impl {
         is_powered_on = false;
         exit_lock = false;
 
+        gpu_core->NotifyShutdown();
+
         services.reset();
         service_manager.reset();
         cheat_engine.reset();

src/core/file_sys/patch_manager.cpp

@@ -128,15 +128,6 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
     if (exefs == nullptr)
         return exefs;
 
-    if (Settings::values.dump_exefs) {
-        LOG_INFO(Loader, "Dumping ExeFS for title_id={:016X}", title_id);
-        const auto dump_dir = fs_controller.GetModificationDumpRoot(title_id);
-        if (dump_dir != nullptr) {
-            const auto exefs_dir = GetOrCreateDirectoryRelative(dump_dir, "/exefs");
-            VfsRawCopyD(exefs, exefs_dir);
-        }
-    }
-
     const auto& disabled = Settings::values.disabled_addons[title_id];
     const auto update_disabled =
         std::find(disabled.cbegin(), disabled.cend(), "Update") != disabled.cend();
@@ -179,6 +170,15 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
         }
     }
 
+    if (Settings::values.dump_exefs) {
+        LOG_INFO(Loader, "Dumping ExeFS for title_id={:016X}", title_id);
+        const auto dump_dir = fs_controller.GetModificationDumpRoot(title_id);
+        if (dump_dir != nullptr) {
+            const auto exefs_dir = GetOrCreateDirectoryRelative(dump_dir, "/exefs");
+            VfsRawCopyD(exefs, exefs_dir);
+        }
+    }
+
     return exefs;
 }
 

src/core/file_sys/vfs.h

@@ -12,6 +12,7 @@
 #include <type_traits>
 #include <vector>
 
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "core/file_sys/vfs_types.h"
 
@@ -29,8 +30,11 @@ enum class VfsEntryType {
 // A class representing an abstract filesystem. A default implementation given the root VirtualDir
 // is provided for convenience, but if the Vfs implementation has any additional state or
 // functionality, they will need to override.
-class VfsFilesystem : NonCopyable {
+class VfsFilesystem {
 public:
+    YUZU_NON_COPYABLE(VfsFilesystem);
+    YUZU_NON_MOVEABLE(VfsFilesystem);
+
     explicit VfsFilesystem(VirtualDir root);
     virtual ~VfsFilesystem();
 
@@ -77,8 +81,12 @@ protected:
 };
 
 // A class representing a file in an abstract filesystem.
-class VfsFile : NonCopyable {
+class VfsFile {
 public:
+    YUZU_NON_COPYABLE(VfsFile);
+    YUZU_NON_MOVEABLE(VfsFile);
+
+    VfsFile() = default;
     virtual ~VfsFile();
 
     // Retrieves the file name.
@@ -176,8 +184,12 @@ public:
 };
 
 // A class representing a directory in an abstract filesystem.
-class VfsDirectory : NonCopyable {
+class VfsDirectory {
 public:
+    YUZU_NON_COPYABLE(VfsDirectory);
+    YUZU_NON_MOVEABLE(VfsDirectory);
+
+    VfsDirectory() = default;
     virtual ~VfsDirectory();
 
     // Retrives the file located at path as if the current directory was root. Returns nullptr if

src/core/frontend/applets/profile_select.cpp

@@ -13,8 +13,7 @@ ProfileSelectApplet::~ProfileSelectApplet() = default;
 void DefaultProfileSelectApplet::SelectProfile(
     std::function<void(std::optional<Common::UUID>)> callback) const {
     Service::Account::ProfileManager manager;
-    callback(manager.GetUser(Settings::values.current_user.GetValue())
-                 .value_or(Common::UUID{Common::INVALID_UUID}));
+    callback(manager.GetUser(Settings::values.current_user.GetValue()).value_or(Common::UUID{}));
     LOG_INFO(Service_ACC, "called, selecting current user instead of prompting...");
 }
 

src/core/hid/emulated_console.cpp

@@ -158,7 +158,7 @@ void EmulatedConsole::SetMotion(const Common::Input::CallbackStatus& callback) {
     auto& motion = console.motion_state;
     motion.accel = emulated.GetAcceleration();
     motion.gyro = emulated.GetGyroscope();
-    motion.rotation = emulated.GetGyroscope();
+    motion.rotation = emulated.GetRotations();
     motion.orientation = emulated.GetOrientation();
     motion.quaternion = emulated.GetQuaternion();
     motion.gyro_bias = emulated.GetGyroBias();

src/core/hid/emulated_console.h

@@ -10,6 +10,7 @@
 #include <mutex>
 #include <unordered_map>
 
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/input.h"
 #include "common/param_package.h"

src/core/hid/emulated_controller.cpp

@@ -145,7 +145,7 @@ void EmulatedController::LoadDevices() {
                    motion_devices.begin(), Common::Input::CreateDevice<Common::Input::InputDevice>);
     std::transform(trigger_params.begin(), trigger_params.end(), trigger_devices.begin(),
                    Common::Input::CreateDevice<Common::Input::InputDevice>);
-    std::transform(battery_params.begin(), battery_params.begin(), battery_devices.end(),
+    std::transform(battery_params.begin(), battery_params.end(), battery_devices.begin(),
                    Common::Input::CreateDevice<Common::Input::InputDevice>);
     std::transform(output_params.begin(), output_params.end(), output_devices.begin(),
                    Common::Input::CreateDevice<Common::Input::OutputDevice>);
@@ -269,7 +269,8 @@ void EmulatedController::ReloadInput() {
     }
 
     // Use a common UUID for TAS
-    const auto tas_uuid = Common::UUID{0x0, 0x7A5};
+    static constexpr Common::UUID TAS_UUID = Common::UUID{
+        {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x7, 0xA5, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}};
 
     // Register TAS devices. No need to force update
     for (std::size_t index = 0; index < tas_button_devices.size(); ++index) {
@@ -278,8 +279,8 @@ void EmulatedController::ReloadInput() {
         }
         tas_button_devices[index]->SetCallback({
             .on_change =
-                [this, index, tas_uuid](const Common::Input::CallbackStatus& callback) {
-                    SetButton(callback, index, tas_uuid);
+                [this, index](const Common::Input::CallbackStatus& callback) {
+                    SetButton(callback, index, TAS_UUID);
                 },
         });
     }
@@ -290,8 +291,8 @@ void EmulatedController::ReloadInput() {
         }
         tas_stick_devices[index]->SetCallback({
             .on_change =
-                [this, index, tas_uuid](const Common::Input::CallbackStatus& callback) {
-                    SetStick(callback, index, tas_uuid);
+                [this, index](const Common::Input::CallbackStatus& callback) {
+                    SetStick(callback, index, TAS_UUID);
                 },
         });
     }
@@ -351,6 +352,19 @@ void EmulatedController::DisableConfiguration() {
     }
 }
 
+void EmulatedController::EnableSystemButtons() {
+    system_buttons_enabled = true;
+}
+
+void EmulatedController::DisableSystemButtons() {
+    system_buttons_enabled = false;
+}
+
+void EmulatedController::ResetSystemButtons() {
+    controller.home_button_state.home.Assign(false);
+    controller.capture_button_state.capture.Assign(false);
+}
+
 bool EmulatedController::IsConfiguring() const {
     return is_configuring;
 }
@@ -389,7 +403,8 @@ std::vector<Common::ParamPackage> EmulatedController::GetMappedDevices(
             devices.begin(), devices.end(), [param](const Common::ParamPackage param_) {
                 return param.Get("engine", "") == param_.Get("engine", "") &&
                        param.Get("guid", "") == param_.Get("guid", "") &&
-                       param.Get("port", 0) == param_.Get("port", 0);
+                       param.Get("port", 0) == param_.Get("port", 0) &&
+                       param.Get("pad", 0) == param_.Get("pad", 0);
             });
         if (devices_it != devices.end()) {
             continue;
@@ -398,6 +413,7 @@ std::vector<Common::ParamPackage> EmulatedController::GetMappedDevices(
         device.Set("engine", param.Get("engine", ""));
         device.Set("guid", param.Get("guid", ""));
         device.Set("port", param.Get("port", 0));
+        device.Set("pad", param.Get("pad", 0));
         devices.push_back(device);
     }
 
@@ -412,7 +428,8 @@ std::vector<Common::ParamPackage> EmulatedController::GetMappedDevices(
             devices.begin(), devices.end(), [param](const Common::ParamPackage param_) {
                 return param.Get("engine", "") == param_.Get("engine", "") &&
                        param.Get("guid", "") == param_.Get("guid", "") &&
-                       param.Get("port", 0) == param_.Get("port", 0);
+                       param.Get("port", 0) == param_.Get("port", 0) &&
+                       param.Get("pad", 0) == param_.Get("pad", 0);
             });
         if (devices_it != devices.end()) {
             continue;
@@ -421,6 +438,7 @@ std::vector<Common::ParamPackage> EmulatedController::GetMappedDevices(
         device.Set("engine", param.Get("engine", ""));
         device.Set("guid", param.Get("guid", ""));
         device.Set("port", param.Get("port", 0));
+        device.Set("pad", param.Get("pad", 0));
         devices.push_back(device);
     }
     return devices;
@@ -596,7 +614,16 @@ void EmulatedController::SetButton(const Common::Input::CallbackStatus& callback
             controller.npad_button_state.right_sr.Assign(current_status.value);
             break;
         case Settings::NativeButton::Home:
+            if (!system_buttons_enabled) {
+                break;
+            }
+            controller.home_button_state.home.Assign(current_status.value);
+            break;
         case Settings::NativeButton::Screenshot:
+            if (!system_buttons_enabled) {
+                break;
+            }
+            controller.capture_button_state.capture.Assign(current_status.value);
             break;
         }
     }
@@ -723,6 +750,7 @@ void EmulatedController::SetMotion(const Common::Input::CallbackStatus& callback
         raw_status.gyro.y.value,
         raw_status.gyro.z.value,
     });
+    emulated.SetGyroThreshold(raw_status.gyro.x.properties.threshold);
     emulated.UpdateRotation(raw_status.delta_timestamp);
     emulated.UpdateOrientation(raw_status.delta_timestamp);
     force_update_motion = raw_status.force_update;
@@ -857,6 +885,12 @@ bool EmulatedController::TestVibration(std::size_t device_index) {
     return SetVibration(device_index, DEFAULT_VIBRATION_VALUE);
 }
 
+bool EmulatedController::SetPollingMode(Common::Input::PollingMode polling_mode) {
+    LOG_INFO(Service_HID, "Set polling mode {}", polling_mode);
+    auto& output_device = output_devices[static_cast<std::size_t>(DeviceIndex::Right)];
+    return output_device->SetPollingMode(polling_mode) == Common::Input::PollingError::None;
+}
+
 void EmulatedController::SetLedPattern() {
     for (auto& device : output_devices) {
         if (!device) {
@@ -879,10 +913,36 @@ void EmulatedController::SetSupportedNpadStyleTag(NpadStyleTag supported_styles)
     if (!is_connected) {
         return;
     }
-    if (!IsControllerSupported()) {
-        LOG_ERROR(Service_HID, "Controller type {} is not supported. Disconnecting controller",
-                  npad_type);
-        Disconnect();
+    if (IsControllerSupported()) {
+        return;
+    }
+
+    Disconnect();
+
+    // Fallback fullkey controllers to Pro controllers
+    if (IsControllerFullkey() && supported_style_tag.fullkey) {
+        LOG_WARNING(Service_HID, "Reconnecting controller type {} as Pro controller", npad_type);
+        SetNpadStyleIndex(NpadStyleIndex::ProController);
+        Connect();
+        return;
+    }
+
+    LOG_ERROR(Service_HID, "Controller type {} is not supported. Disconnecting controller",
+              npad_type);
+}
+
+bool EmulatedController::IsControllerFullkey(bool use_temporary_value) const {
+    const auto type = is_configuring && use_temporary_value ? tmp_npad_type : npad_type;
+    switch (type) {
+    case NpadStyleIndex::ProController:
+    case NpadStyleIndex::GameCube:
+    case NpadStyleIndex::NES:
+    case NpadStyleIndex::SNES:
+    case NpadStyleIndex::N64:
+    case NpadStyleIndex::SegaGenesis:
+        return true;
+    default:
+        return false;
     }
 }
 
@@ -1051,6 +1111,20 @@ BatteryValues EmulatedController::GetBatteryValues() const {
     return controller.battery_values;
 }
 
+HomeButtonState EmulatedController::GetHomeButtons() const {
+    if (is_configuring) {
+        return {};
+    }
+    return controller.home_button_state;
+}
+
+CaptureButtonState EmulatedController::GetCaptureButtons() const {
+    if (is_configuring) {
+        return {};
+    }
+    return controller.capture_button_state;
+}
+
 NpadButtonState EmulatedController::GetNpadButtons() const {
     if (is_configuring) {
         return {};

src/core/hid/emulated_controller.h

@@ -13,8 +13,6 @@
 #include "common/common_types.h"
 #include "common/input.h"
 #include "common/param_package.h"
-#include "common/point.h"
-#include "common/quaternion.h"
 #include "common/settings.h"
 #include "common/vector_math.h"
 #include "core/hid/hid_types.h"
@@ -101,6 +99,8 @@ struct ControllerStatus {
     VibrationValues vibration_values{};
 
     // Data for HID serices
+    HomeButtonState home_button_state{};
+    CaptureButtonState capture_button_state{};
     NpadButtonState npad_button_state{};
     DebugPadButton debug_pad_button_state{};
     AnalogSticks analog_stick_state{};
@@ -198,6 +198,15 @@ public:
     /// Returns the emulated controller into normal mode, allowing the modification of the HID state
     void DisableConfiguration();
 
+    /// Enables Home and Screenshot buttons
+    void EnableSystemButtons();
+
+    /// Disables Home and Screenshot buttons
+    void DisableSystemButtons();
+
+    /// Sets Home and Screenshot buttons to false
+    void ResetSystemButtons();
+
     /// Returns true if the emulated controller is in configuring mode
     bool IsConfiguring() const;
 
@@ -261,7 +270,13 @@ public:
     /// Returns the latest battery status from the controller with parameters
     BatteryValues GetBatteryValues() const;
 
-    /// Returns the latest status of button input for the npad service
+    /// Returns the latest status of button input for the hid::HomeButton service
+    HomeButtonState GetHomeButtons() const;
+
+    /// Returns the latest status of button input for the hid::CaptureButton service
+    CaptureButtonState GetCaptureButtons() const;
+
+    /// Returns the latest status of button input for the hid::Npad service
     NpadButtonState GetNpadButtons() const;
 
     /// Returns the latest status of button input for the debug pad service
@@ -284,16 +299,23 @@ public:
 
     /**
      * Sends a specific vibration to the output device
-     * @return returns true if vibration had no errors
+     * @return true if vibration had no errors
      */
     bool SetVibration(std::size_t device_index, VibrationValue vibration);
 
     /**
      * Sends a small vibration to the output device
-     * @return returns true if SetVibration was successfull
+     * @return true if SetVibration was successfull
      */
     bool TestVibration(std::size_t device_index);
 
+    /**
+     * Sets the desired data to be polled from a controller
+     * @param polling_mode type of input desired buttons, gyro, nfc, ir, etc.
+     * @return true if SetPollingMode was successfull
+     */
+    bool SetPollingMode(Common::Input::PollingMode polling_mode);
+
     /// Returns the led pattern corresponding to this emulated controller
     LedPattern GetLedPattern() const;
 
@@ -320,6 +342,12 @@ private:
     /// Set the params for TAS devices
     void LoadTASParams();
 
+    /**
+     * @param use_temporary_value If true tmp_npad_type will be used
+     * @return true if the controller style is fullkey
+     */
+    bool IsControllerFullkey(bool use_temporary_value = false) const;
+
     /**
      * Checks the current controller type against the supported_style_tag
      * @param use_temporary_value If true tmp_npad_type will be used
@@ -377,6 +405,7 @@ private:
     NpadStyleTag supported_style_tag{NpadStyleSet::All};
     bool is_connected{false};
     bool is_configuring{false};
+    bool system_buttons_enabled{true};
     f32 motion_sensitivity{0.01f};
     bool force_update_motion{false};
 

src/core/hid/hid_core.h

@@ -6,6 +6,7 @@
 
 #include <memory>
 
+#include "common/common_funcs.h"
 #include "core/hid/hid_types.h"
 
 namespace Core::HID {

src/core/hid/hid_types.h

@@ -378,6 +378,26 @@ struct LedPattern {
     };
 };
 
+struct HomeButtonState {
+    union {
+        u64 raw{};
+
+        // Buttons
+        BitField<0, 1, u64> home;
+    };
+};
+static_assert(sizeof(HomeButtonState) == 0x8, "HomeButtonState has incorrect size.");
+
+struct CaptureButtonState {
+    union {
+        u64 raw{};
+
+        // Buttons
+        BitField<0, 1, u64> capture;
+    };
+};
+static_assert(sizeof(CaptureButtonState) == 0x8, "CaptureButtonState has incorrect size.");
+
 struct NpadButtonState {
     union {
         NpadButton raw{};

src/core/hid/input_converter.cpp

@@ -28,7 +28,7 @@ Common::Input::BatteryStatus TransformToBattery(const Common::Input::CallbackSta
         if (value > 0.8f) {
             battery = Common::Input::BatteryLevel::Full;
         }
-        if (value >= 1.0f) {
+        if (value >= 0.95f) {
             battery = Common::Input::BatteryLevel::Charging;
         }
         break;
@@ -114,7 +114,7 @@ Common::Input::MotionStatus TransformToMotion(const Common::Input::CallbackStatu
         if (TransformToButton(callback).value) {
             std::random_device device;
             std::mt19937 gen(device());
-            std::uniform_int_distribution<s16> distribution(-1000, 1000);
+            std::uniform_int_distribution<s16> distribution(-5000, 5000);
             status.accel.x.raw_value = static_cast<f32>(distribution(gen)) * 0.001f;
             status.accel.y.raw_value = static_cast<f32>(distribution(gen)) * 0.001f;
             status.accel.z.raw_value = static_cast<f32>(distribution(gen)) * 0.001f;

src/core/hid/motion_input.cpp

@@ -10,6 +10,7 @@ namespace Core::HID {
 MotionInput::MotionInput() {
     // Initialize PID constants with default values
     SetPID(0.3f, 0.005f, 0.0f);
+    SetGyroThreshold(0.007f);
 }
 
 void MotionInput::SetPID(f32 new_kp, f32 new_ki, f32 new_kd) {
@@ -30,7 +31,7 @@ void MotionInput::SetGyroscope(const Common::Vec3f& gyroscope) {
         gyro_bias = (gyro_bias * 0.9999f) + (gyroscope * 0.0001f);
     }
 
-    if (gyro.Length2() < gyro_threshold) {
+    if (gyro.Length() < gyro_threshold) {
         gyro = {};
     } else {
         only_accelerometer = false;

src/core/hle/ipc_helpers.h

@@ -404,6 +404,11 @@ inline s32 RequestParser::Pop() {
     return static_cast<s32>(Pop<u32>());
 }
 
+// Ignore the -Wclass-memaccess warning on memcpy for non-trivially default constructible objects.
+#if defined(__GNUC__)
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wclass-memaccess"
+#endif
 template <typename T>
 void RequestParser::PopRaw(T& value) {
     static_assert(std::is_trivially_copyable_v<T>,
@@ -411,6 +416,9 @@ void RequestParser::PopRaw(T& value) {
     std::memcpy(&value, cmdbuf + index, sizeof(T));
     index += (sizeof(T) + 3) / 4; // round up to word length
 }
+#if defined(__GNUC__)
+#pragma GCC diagnostic pop
+#endif
 
 template <typename T>
 T RequestParser::PopRaw() {

src/core/hle/kernel/global_scheduler_context.cpp

@@ -9,6 +9,7 @@
 #include "core/hle/kernel/global_scheduler_context.h"
 #include "core/hle/kernel/k_scheduler.h"
 #include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/physical_core.h"
 
 namespace Kernel {
 
@@ -42,6 +43,11 @@ void GlobalSchedulerContext::PreemptThreads() {
     for (u32 core_id = 0; core_id < Core::Hardware::NUM_CPU_CORES; core_id++) {
         const u32 priority = preemption_priorities[core_id];
         kernel.Scheduler(core_id).RotateScheduledQueue(core_id, priority);
+
+        // Signal an interrupt occurred. For core 3, this is a certainty, as preemption will result
+        // in the rotator thread being scheduled. For cores 0-2, this is to simulate or system
+        // interrupts that may have occurred.
+        kernel.PhysicalCore(core_id).Interrupt();
     }
 }
 

src/core/hle/kernel/hle_ipc.h

@@ -341,10 +341,6 @@ public:
         return *thread;
     }
 
-    bool IsThreadWaiting() const {
-        return is_thread_waiting;
-    }
-
 private:
     friend class IPC::ResponseBuilder;
 
@@ -379,7 +375,6 @@ private:
     u32 domain_offset{};
 
     std::shared_ptr<SessionRequestManager> manager;
-    bool is_thread_waiting{};
 
     KernelCore& kernel;
     Core::Memory::Memory& memory;

src/core/hle/kernel/k_affinity_mask.h

@@ -31,8 +31,6 @@ public:
     }
 
     constexpr void SetAffinity(s32 core, bool set) {
-        ASSERT(0 <= core && core < static_cast<s32>(Core::Hardware::NUM_CPU_CORES));
-
         if (set) {
             this->mask |= GetCoreBit(core);
         } else {

src/core/hle/kernel/k_auto_object.h

@@ -20,8 +20,6 @@ class KernelCore;
 class KProcess;
 
 #define KERNEL_AUTOOBJECT_TRAITS(CLASS, BASE_CLASS)                                                \
-    YUZU_NON_COPYABLE(CLASS);                                                                      \
-    YUZU_NON_MOVEABLE(CLASS);                                                                      \
                                                                                                    \
 private:                                                                                           \
     friend class ::Kernel::KClassTokenGenerator;                                                   \
@@ -32,6 +30,9 @@ private:
     }                                                                                              \
                                                                                                    \
 public:                                                                                            \
+    YUZU_NON_COPYABLE(CLASS);                                                                      \
+    YUZU_NON_MOVEABLE(CLASS);                                                                      \
+                                                                                                   \
     using BaseClass = BASE_CLASS;                                                                  \
     static constexpr TypeObj GetStaticTypeObj() {                                                  \
         constexpr ClassTokenType Token = ClassToken();                                             \
@@ -224,9 +225,9 @@ private:
 
 template <typename T>
 class KScopedAutoObject {
+public:
     YUZU_NON_COPYABLE(KScopedAutoObject);
 
-public:
     constexpr KScopedAutoObject() = default;
 
     constexpr KScopedAutoObject(T* o) : m_obj(o) {

src/core/hle/kernel/k_auto_object_container.h

@@ -16,13 +16,12 @@ class KernelCore;
 class KProcess;
 
 class KAutoObjectWithListContainer {
+public:
     YUZU_NON_COPYABLE(KAutoObjectWithListContainer);
     YUZU_NON_MOVEABLE(KAutoObjectWithListContainer);
 
-public:
     using ListType = boost::intrusive::rbtree<KAutoObjectWithList>;
 
-public:
     class ListAccessor : public KScopedLightLock {
     public:
         explicit ListAccessor(KAutoObjectWithListContainer* container)
@@ -48,7 +47,6 @@ public:
 
     friend class ListAccessor;
 
-public:
     KAutoObjectWithListContainer(KernelCore& kernel) : m_lock(kernel), m_object_list() {}
 
     void Initialize() {}

src/core/hle/kernel/k_code_memory.cpp

@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include "common/alignment.h"
 #include "common/common_types.h"
 #include "core/device_memory.h"
 #include "core/hle/kernel/k_auto_object.h"
@@ -28,8 +29,7 @@ ResultCode KCodeMemory::Initialize(Core::DeviceMemory& device_memory, VAddr addr
     auto& page_table = m_owner->PageTable();
 
     // Construct the page group.
-    KMemoryInfo kBlockInfo = page_table.QueryInfo(addr);
-    m_page_group = KPageLinkedList(kBlockInfo.GetAddress(), kBlockInfo.GetNumPages());
+    m_page_group = KPageLinkedList(addr, Common::DivideUp(size, PageSize));
 
     // Lock the memory.
     R_TRY(page_table.LockForCodeMemory(addr, size))
@@ -143,4 +143,4 @@ ResultCode KCodeMemory::UnmapFromOwner(VAddr address, size_t size) {
     return ResultSuccess;
 }
 
-} // namespace Kernel
+} // namespace Kernel

src/core/hle/kernel/k_handle_table.h

@@ -22,13 +22,12 @@ namespace Kernel {
 class KernelCore;
 
 class KHandleTable {
+public:
     YUZU_NON_COPYABLE(KHandleTable);
     YUZU_NON_MOVEABLE(KHandleTable);
 
-public:
     static constexpr size_t MaxTableSize = 1024;
 
-public:
     explicit KHandleTable(KernelCore& kernel_);
     ~KHandleTable();
 

src/core/hle/kernel/k_interrupt_manager.cpp

@@ -0,0 +1,34 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/hle/kernel/k_interrupt_manager.h"
+#include "core/hle/kernel/k_process.h"
+#include "core/hle/kernel/k_scheduler.h"
+#include "core/hle/kernel/k_thread.h"
+#include "core/hle/kernel/kernel.h"
+
+namespace Kernel::KInterruptManager {
+
+void HandleInterrupt(KernelCore& kernel, s32 core_id) {
+    auto* process = kernel.CurrentProcess();
+    if (!process) {
+        return;
+    }
+
+    auto& scheduler = kernel.Scheduler(core_id);
+    auto& current_thread = *scheduler.GetCurrentThread();
+
+    // If the user disable count is set, we may need to pin the current thread.
+    if (current_thread.GetUserDisableCount() && !process->GetPinnedThread(core_id)) {
+        KScopedSchedulerLock sl{kernel};
+
+        // Pin the current thread.
+        process->PinCurrentThread(core_id);
+
+        // Set the interrupt flag for the thread.
+        scheduler.GetCurrentThread()->SetInterruptFlag();
+    }
+}
+
+} // namespace Kernel::KInterruptManager

src/core/hle/kernel/k_interrupt_manager.h

@@ -0,0 +1,17 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+
+namespace Kernel {
+
+class KernelCore;
+
+namespace KInterruptManager {
+void HandleInterrupt(KernelCore& kernel, s32 core_id);
+}
+
+} // namespace Kernel

src/core/hle/kernel/k_memory_block.h

@@ -70,12 +70,12 @@ enum class KMemoryState : u32 {
     ThreadLocal =
         static_cast<u32>(Svc::MemoryState::ThreadLocal) | FlagMapped | FlagReferenceCounted,
 
-    Transferred = static_cast<u32>(Svc::MemoryState::Transferred) | FlagsMisc |
-                  FlagCanAlignedDeviceMap | FlagCanChangeAttribute | FlagCanUseIpc |
-                  FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc,
+    Transfered = static_cast<u32>(Svc::MemoryState::Transferred) | FlagsMisc |
+                 FlagCanAlignedDeviceMap | FlagCanChangeAttribute | FlagCanUseIpc |
+                 FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc,
 
-    SharedTransferred = static_cast<u32>(Svc::MemoryState::SharedTransferred) | FlagsMisc |
-                        FlagCanAlignedDeviceMap | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc,
+    SharedTransfered = static_cast<u32>(Svc::MemoryState::SharedTransferred) | FlagsMisc |
+                       FlagCanAlignedDeviceMap | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc,
 
     SharedCode = static_cast<u32>(Svc::MemoryState::SharedCode) | FlagMapped |
                  FlagReferenceCounted | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc,
@@ -93,6 +93,8 @@ enum class KMemoryState : u32 {
     GeneratedCode = static_cast<u32>(Svc::MemoryState::GeneratedCode) | FlagMapped |
                     FlagReferenceCounted | FlagCanDebug,
     CodeOut = static_cast<u32>(Svc::MemoryState::CodeOut) | FlagMapped | FlagReferenceCounted,
+
+    Coverage = static_cast<u32>(Svc::MemoryState::Coverage) | FlagMapped,
 };
 DECLARE_ENUM_FLAG_OPERATORS(KMemoryState);
 
@@ -108,8 +110,8 @@ static_assert(static_cast<u32>(KMemoryState::AliasCodeData) == 0x03FFBD09);
 static_assert(static_cast<u32>(KMemoryState::Ipc) == 0x005C3C0A);
 static_assert(static_cast<u32>(KMemoryState::Stack) == 0x005C3C0B);
 static_assert(static_cast<u32>(KMemoryState::ThreadLocal) == 0x0040200C);
-static_assert(static_cast<u32>(KMemoryState::Transferred) == 0x015C3C0D);
-static_assert(static_cast<u32>(KMemoryState::SharedTransferred) == 0x005C380E);
+static_assert(static_cast<u32>(KMemoryState::Transfered) == 0x015C3C0D);
+static_assert(static_cast<u32>(KMemoryState::SharedTransfered) == 0x005C380E);
 static_assert(static_cast<u32>(KMemoryState::SharedCode) == 0x0040380F);
 static_assert(static_cast<u32>(KMemoryState::Inaccessible) == 0x00000010);
 static_assert(static_cast<u32>(KMemoryState::NonSecureIpc) == 0x005C3811);
@@ -117,6 +119,7 @@ static_assert(static_cast<u32>(KMemoryState::NonDeviceIpc) == 0x004C2812);
 static_assert(static_cast<u32>(KMemoryState::Kernel) == 0x00002013);
 static_assert(static_cast<u32>(KMemoryState::GeneratedCode) == 0x00402214);
 static_assert(static_cast<u32>(KMemoryState::CodeOut) == 0x00402015);
+static_assert(static_cast<u32>(KMemoryState::Coverage) == 0x00002016);
 
 enum class KMemoryPermission : u8 {
     None = 0,
@@ -155,7 +158,13 @@ enum class KMemoryPermission : u8 {
 DECLARE_ENUM_FLAG_OPERATORS(KMemoryPermission);
 
 constexpr KMemoryPermission ConvertToKMemoryPermission(Svc::MemoryPermission perm) {
-    return static_cast<KMemoryPermission>(perm);
+    return static_cast<KMemoryPermission>(
+        (static_cast<KMemoryPermission>(perm) & KMemoryPermission::UserMask) |
+        KMemoryPermission::KernelRead |
+        ((static_cast<KMemoryPermission>(perm) & KMemoryPermission::UserWrite)
+         << KMemoryPermission::KernelShift) |
+        (perm == Svc::MemoryPermission::None ? KMemoryPermission::NotMapped
+                                             : KMemoryPermission::None));
 }
 
 enum class KMemoryAttribute : u8 {
@@ -169,6 +178,8 @@ enum class KMemoryAttribute : u8 {
     DeviceShared = static_cast<u8>(Svc::MemoryAttribute::DeviceShared),
     Uncached = static_cast<u8>(Svc::MemoryAttribute::Uncached),
 
+    SetMask = Uncached,
+
     IpcAndDeviceMapped = IpcLocked | DeviceShared,
     LockedAndIpcLocked = Locked | IpcLocked,
     DeviceSharedAndUncached = DeviceShared | Uncached
@@ -215,6 +226,15 @@ struct KMemoryInfo {
     constexpr VAddr GetLastAddress() const {
         return GetEndAddress() - 1;
     }
+    constexpr KMemoryState GetState() const {
+        return state;
+    }
+    constexpr KMemoryAttribute GetAttribute() const {
+        return attribute;
+    }
+    constexpr KMemoryPermission GetPermission() const {
+        return perm;
+    }
 };
 
 class KMemoryBlock final {

src/core/hle/kernel/k_memory_manager.cpp

@@ -8,12 +8,16 @@
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "common/scope_exit.h"
+#include "core/core.h"
+#include "core/device_memory.h"
 #include "core/hle/kernel/k_memory_manager.h"
 #include "core/hle/kernel/k_page_linked_list.h"
 #include "core/hle/kernel/svc_results.h"
 
 namespace Kernel {
 
+KMemoryManager::KMemoryManager(Core::System& system_) : system{system_} {}
+
 std::size_t KMemoryManager::Impl::Initialize(Pool new_pool, u64 start_address, u64 end_address) {
     const auto size{end_address - start_address};
 
@@ -81,7 +85,7 @@ VAddr KMemoryManager::AllocateAndOpenContinuous(std::size_t num_pages, std::size
 }
 
 ResultCode KMemoryManager::Allocate(KPageLinkedList& page_list, std::size_t num_pages, Pool pool,
-                                    Direction dir) {
+                                    Direction dir, u32 heap_fill_value) {
     ASSERT(page_list.GetNumPages() == 0);
 
     // Early return if we're allocating no pages
@@ -139,6 +143,12 @@ ResultCode KMemoryManager::Allocate(KPageLinkedList& page_list, std::size_t num_
         }
     }
 
+    // Clear allocated memory.
+    for (const auto& it : page_list.Nodes()) {
+        std::memset(system.DeviceMemory().GetPointer(it.GetAddress()), heap_fill_value,
+                    it.GetSize());
+    }
+
     // Only succeed if we allocated as many pages as we wanted
     if (num_pages) {
         return ResultOutOfMemory;
@@ -146,11 +156,12 @@ ResultCode KMemoryManager::Allocate(KPageLinkedList& page_list, std::size_t num_
 
     // We succeeded!
     group_guard.Cancel();
+
     return ResultSuccess;
 }
 
 ResultCode KMemoryManager::Free(KPageLinkedList& page_list, std::size_t num_pages, Pool pool,
-                                Direction dir) {
+                                Direction dir, u32 heap_fill_value) {
     // Early return if we're freeing no pages
     if (!num_pages) {
         return ResultSuccess;

src/core/hle/kernel/k_memory_manager.h

@@ -8,16 +8,24 @@
 #include <mutex>
 #include <tuple>
 
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "core/hle/kernel/k_page_heap.h"
 #include "core/hle/result.h"
 
+namespace Core {
+class System;
+}
+
 namespace Kernel {
 
 class KPageLinkedList;
 
-class KMemoryManager final : NonCopyable {
+class KMemoryManager final {
 public:
+    YUZU_NON_COPYABLE(KMemoryManager);
+    YUZU_NON_MOVEABLE(KMemoryManager);
+
     enum class Pool : u32 {
         Application = 0,
         Applet = 1,
@@ -42,7 +50,7 @@ public:
         Mask = (0xF << Shift),
     };
 
-    KMemoryManager() = default;
+    explicit KMemoryManager(Core::System& system_);
 
     constexpr std::size_t GetSize(Pool pool) const {
         return managers[static_cast<std::size_t>(pool)].GetSize();
@@ -51,10 +59,10 @@ public:
     void InitializeManager(Pool pool, u64 start_address, u64 end_address);
 
     VAddr AllocateAndOpenContinuous(size_t num_pages, size_t align_pages, u32 option);
-    ResultCode Allocate(KPageLinkedList& page_list, std::size_t num_pages, Pool pool,
-                        Direction dir = Direction::FromFront);
-    ResultCode Free(KPageLinkedList& page_list, std::size_t num_pages, Pool pool,
-                    Direction dir = Direction::FromFront);
+    ResultCode Allocate(KPageLinkedList& page_list, std::size_t num_pages, Pool pool, Direction dir,
+                        u32 heap_fill_value = 0);
+    ResultCode Free(KPageLinkedList& page_list, std::size_t num_pages, Pool pool, Direction dir,
+                    u32 heap_fill_value = 0);
 
     static constexpr std::size_t MaxManagerCount = 10;
 
@@ -84,26 +92,13 @@ public:
     }
 
 private:
-    class Impl final : NonCopyable {
-    private:
-        using RefCount = u16;
-
-    private:
-        KPageHeap heap;
-        Pool pool{};
-
+    class Impl final {
     public:
-        static std::size_t CalculateManagementOverheadSize(std::size_t region_size);
+        YUZU_NON_COPYABLE(Impl);
+        YUZU_NON_MOVEABLE(Impl);
 
-        static constexpr std::size_t CalculateOptimizedProcessOverheadSize(
-            std::size_t region_size) {
-            return (Common::AlignUp((region_size / PageSize), Common::BitSize<u64>()) /
-                    Common::BitSize<u64>()) *
-                   sizeof(u64);
-        }
-
-    public:
         Impl() = default;
+        ~Impl() = default;
 
         std::size_t Initialize(Pool new_pool, u64 start_address, u64 end_address);
 
@@ -126,9 +121,25 @@ private:
         constexpr VAddr GetEndAddress() const {
             return heap.GetEndAddress();
         }
+
+        static std::size_t CalculateManagementOverheadSize(std::size_t region_size);
+
+        static constexpr std::size_t CalculateOptimizedProcessOverheadSize(
+            std::size_t region_size) {
+            return (Common::AlignUp((region_size / PageSize), Common::BitSize<u64>()) /
+                    Common::BitSize<u64>()) *
+                   sizeof(u64);
+        }
+
+    private:
+        using RefCount = u16;
+
+        KPageHeap heap;
+        Pool pool{};
     };
 
 private:
+    Core::System& system;
     std::array<std::mutex, static_cast<std::size_t>(Pool::Count)> pool_locks;
     std::array<Impl, MaxManagerCount> managers;
 };

src/core/hle/kernel/k_memory_region.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include "common/assert.h"
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/intrusive_red_black_tree.h"
 #include "core/hle/kernel/k_memory_region_type.h"
@@ -13,11 +14,13 @@ namespace Kernel {
 
 class KMemoryRegionAllocator;
 
-class KMemoryRegion final : public Common::IntrusiveRedBlackTreeBaseNode<KMemoryRegion>,
-                            NonCopyable {
+class KMemoryRegion final : public Common::IntrusiveRedBlackTreeBaseNode<KMemoryRegion> {
     friend class KMemoryRegionTree;
 
 public:
+    YUZU_NON_COPYABLE(KMemoryRegion);
+    YUZU_NON_MOVEABLE(KMemoryRegion);
+
     constexpr KMemoryRegion() = default;
     constexpr KMemoryRegion(u64 address_, u64 last_address_)
         : address{address_}, last_address{last_address_} {}
@@ -29,6 +32,8 @@ public:
         : KMemoryRegion(address_, last_address_, std::numeric_limits<u64>::max(), attributes_,
                         type_id_) {}
 
+    ~KMemoryRegion() = default;
+
     static constexpr int Compare(const KMemoryRegion& lhs, const KMemoryRegion& rhs) {
         if (lhs.GetAddress() < rhs.GetAddress()) {
             return -1;
@@ -39,16 +44,6 @@ public:
         }
     }
 
-private:
-    constexpr void Reset(u64 a, u64 la, u64 p, u32 r, u32 t) {
-        address = a;
-        pair_address = p;
-        last_address = la;
-        attributes = r;
-        type_id = t;
-    }
-
-public:
     constexpr u64 GetAddress() const {
         return address;
     }
@@ -108,6 +103,14 @@ public:
     }
 
 private:
+    constexpr void Reset(u64 a, u64 la, u64 p, u32 r, u32 t) {
+        address = a;
+        pair_address = p;
+        last_address = la;
+        attributes = r;
+        type_id = t;
+    }
+
     u64 address{};
     u64 last_address{};
     u64 pair_address{};
@@ -115,8 +118,25 @@ private:
     u32 type_id{};
 };
 
-class KMemoryRegionTree final : NonCopyable {
+class KMemoryRegionTree final {
+private:
+    using TreeType =
+        Common::IntrusiveRedBlackTreeBaseTraits<KMemoryRegion>::TreeType<KMemoryRegion>;
+
 public:
+    YUZU_NON_COPYABLE(KMemoryRegionTree);
+    YUZU_NON_MOVEABLE(KMemoryRegionTree);
+
+    using value_type = TreeType::value_type;
+    using size_type = TreeType::size_type;
+    using difference_type = TreeType::difference_type;
+    using pointer = TreeType::pointer;
+    using const_pointer = TreeType::const_pointer;
+    using reference = TreeType::reference;
+    using const_reference = TreeType::const_reference;
+    using iterator = TreeType::iterator;
+    using const_iterator = TreeType::const_iterator;
+
     struct DerivedRegionExtents {
         const KMemoryRegion* first_region{};
         const KMemoryRegion* last_region{};
@@ -140,29 +160,9 @@ public:
         }
     };
 
-private:
-    using TreeType =
-        Common::IntrusiveRedBlackTreeBaseTraits<KMemoryRegion>::TreeType<KMemoryRegion>;
-
-public:
-    using value_type = TreeType::value_type;
-    using size_type = TreeType::size_type;
-    using difference_type = TreeType::difference_type;
-    using pointer = TreeType::pointer;
-    using const_pointer = TreeType::const_pointer;
-    using reference = TreeType::reference;
-    using const_reference = TreeType::const_reference;
-    using iterator = TreeType::iterator;
-    using const_iterator = TreeType::const_iterator;
-
-private:
-    TreeType m_tree{};
-    KMemoryRegionAllocator& memory_region_allocator;
-
-public:
     explicit KMemoryRegionTree(KMemoryRegionAllocator& memory_region_allocator_);
+    ~KMemoryRegionTree() = default;
 
-public:
     KMemoryRegion* FindModifiable(u64 address) {
         if (auto it = this->find(KMemoryRegion(address, address, 0, 0)); it != this->end()) {
             return std::addressof(*it);
@@ -241,7 +241,6 @@ public:
         return GetDerivedRegionExtents(static_cast<KMemoryRegionType>(type_id));
     }
 
-public:
     void InsertDirectly(u64 address, u64 last_address, u32 attr = 0, u32 type_id = 0);
     bool Insert(u64 address, size_t size, u32 type_id, u32 new_attr = 0, u32 old_attr = 0);
 
@@ -252,7 +251,6 @@ public:
         return this->GetRandomAlignedRegion(size + 2 * guard_size, alignment, type_id) + guard_size;
     }
 
-public:
     // Iterator accessors.
     iterator begin() {
         return m_tree.begin();
@@ -322,13 +320,21 @@ public:
     iterator nfind(const_reference ref) const {
         return m_tree.nfind(ref);
     }
+
+private:
+    TreeType m_tree{};
+    KMemoryRegionAllocator& memory_region_allocator;
 };
 
-class KMemoryRegionAllocator final : NonCopyable {
+class KMemoryRegionAllocator final {
 public:
+    YUZU_NON_COPYABLE(KMemoryRegionAllocator);
+    YUZU_NON_MOVEABLE(KMemoryRegionAllocator);
+
     static constexpr size_t MaxMemoryRegions = 200;
 
     constexpr KMemoryRegionAllocator() = default;
+    constexpr ~KMemoryRegionAllocator() = default;
 
     template <typename... Args>
     KMemoryRegion* Allocate(Args&&... args) {

src/core/hle/kernel/k_page_heap.h

@@ -8,14 +8,44 @@
 #include <vector>
 
 #include "common/alignment.h"
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "core/hle/kernel/k_page_bitmap.h"
 #include "core/hle/kernel/memory_types.h"
 
 namespace Kernel {
 
-class KPageHeap final : NonCopyable {
+class KPageHeap final {
 public:
+    YUZU_NON_COPYABLE(KPageHeap);
+    YUZU_NON_MOVEABLE(KPageHeap);
+
+    KPageHeap() = default;
+    ~KPageHeap() = default;
+
+    constexpr VAddr GetAddress() const {
+        return heap_address;
+    }
+    constexpr std::size_t GetSize() const {
+        return heap_size;
+    }
+    constexpr VAddr GetEndAddress() const {
+        return GetAddress() + GetSize();
+    }
+    constexpr std::size_t GetPageOffset(VAddr block) const {
+        return (block - GetAddress()) / PageSize;
+    }
+
+    void Initialize(VAddr heap_address, std::size_t heap_size, std::size_t metadata_size);
+    VAddr AllocateBlock(s32 index, bool random);
+    void Free(VAddr addr, std::size_t num_pages);
+
+    void UpdateUsedSize() {
+        used_size = heap_size - (GetNumFreePages() * PageSize);
+    }
+
+    static std::size_t CalculateManagementOverheadSize(std::size_t region_size);
+
     static constexpr s32 GetAlignedBlockIndex(std::size_t num_pages, std::size_t align_pages) {
         const auto target_pages{std::max(num_pages, align_pages)};
         for (std::size_t i = 0; i < NumMemoryBlockPageShifts; i++) {
@@ -45,21 +75,13 @@ public:
     }
 
 private:
-    static constexpr std::size_t NumMemoryBlockPageShifts{7};
-    static constexpr std::array<std::size_t, NumMemoryBlockPageShifts> MemoryBlockPageShifts{
-        0xC, 0x10, 0x15, 0x16, 0x19, 0x1D, 0x1E,
-    };
-
-    class Block final : NonCopyable {
-    private:
-        KPageBitmap bitmap;
-        VAddr heap_address{};
-        uintptr_t end_offset{};
-        std::size_t block_shift{};
-        std::size_t next_block_shift{};
-
+    class Block final {
     public:
+        YUZU_NON_COPYABLE(Block);
+        YUZU_NON_MOVEABLE(Block);
+
         Block() = default;
+        ~Block() = default;
 
         constexpr std::size_t GetShift() const {
             return block_shift;
@@ -129,7 +151,6 @@ private:
             return heap_address + (offset << GetShift());
         }
 
-    public:
         static constexpr std::size_t CalculateManagementOverheadSize(std::size_t region_size,
                                                                      std::size_t cur_block_shift,
                                                                      std::size_t next_block_shift) {
@@ -139,35 +160,15 @@ private:
             return KPageBitmap::CalculateManagementOverheadSize(
                 (align * 2 + Common::AlignUp(region_size, align)) / cur_block_size);
         }
+
+    private:
+        KPageBitmap bitmap;
+        VAddr heap_address{};
+        uintptr_t end_offset{};
+        std::size_t block_shift{};
+        std::size_t next_block_shift{};
     };
 
-public:
-    KPageHeap() = default;
-
-    constexpr VAddr GetAddress() const {
-        return heap_address;
-    }
-    constexpr std::size_t GetSize() const {
-        return heap_size;
-    }
-    constexpr VAddr GetEndAddress() const {
-        return GetAddress() + GetSize();
-    }
-    constexpr std::size_t GetPageOffset(VAddr block) const {
-        return (block - GetAddress()) / PageSize;
-    }
-
-    void Initialize(VAddr heap_address, std::size_t heap_size, std::size_t metadata_size);
-    VAddr AllocateBlock(s32 index, bool random);
-    void Free(VAddr addr, std::size_t num_pages);
-
-    void UpdateUsedSize() {
-        used_size = heap_size - (GetNumFreePages() * PageSize);
-    }
-
-    static std::size_t CalculateManagementOverheadSize(std::size_t region_size);
-
-private:
     constexpr std::size_t GetNumFreePages() const {
         std::size_t num_free{};
 
@@ -180,6 +181,11 @@ private:
 
     void FreeBlock(VAddr block, s32 index);
 
+    static constexpr std::size_t NumMemoryBlockPageShifts{7};
+    static constexpr std::array<std::size_t, NumMemoryBlockPageShifts> MemoryBlockPageShifts{
+        0xC, 0x10, 0x15, 0x16, 0x19, 0x1D, 0x1E,
+    };
+
     VAddr heap_address{};
     std::size_t heap_size{};
     std::size_t used_size{};

src/core/hle/kernel/k_page_table.h

@@ -5,11 +5,12 @@
 #pragma once
 
 #include <memory>
-#include <mutex>
 
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/page_table.h"
 #include "core/file_sys/program_metadata.h"
+#include "core/hle/kernel/k_light_lock.h"
 #include "core/hle/kernel/k_memory_block.h"
 #include "core/hle/kernel/k_memory_manager.h"
 #include "core/hle/result.h"
@@ -22,34 +23,37 @@ namespace Kernel {
 
 class KMemoryBlockManager;
 
-class KPageTable final : NonCopyable {
+class KPageTable final {
 public:
+    YUZU_NON_COPYABLE(KPageTable);
+    YUZU_NON_MOVEABLE(KPageTable);
+
     explicit KPageTable(Core::System& system_);
+    ~KPageTable();
 
     ResultCode InitializeForProcess(FileSys::ProgramAddressSpaceType as_type, bool enable_aslr,
                                     VAddr code_addr, std::size_t code_size,
                                     KMemoryManager::Pool pool);
     ResultCode MapProcessCode(VAddr addr, std::size_t pages_count, KMemoryState state,
                               KMemoryPermission perm);
-    ResultCode MapProcessCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
-    ResultCode UnmapProcessCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
+    ResultCode MapCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
+    ResultCode UnmapCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
     ResultCode UnmapProcessMemory(VAddr dst_addr, std::size_t size, KPageTable& src_page_table,
                                   VAddr src_addr);
     ResultCode MapPhysicalMemory(VAddr addr, std::size_t size);
     ResultCode UnmapPhysicalMemory(VAddr addr, std::size_t size);
-    ResultCode UnmapMemory(VAddr addr, std::size_t size);
-    ResultCode Map(VAddr dst_addr, VAddr src_addr, std::size_t size);
-    ResultCode Unmap(VAddr dst_addr, VAddr src_addr, std::size_t size);
+    ResultCode MapMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
+    ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
     ResultCode MapPages(VAddr addr, KPageLinkedList& page_linked_list, KMemoryState state,
                         KMemoryPermission perm);
     ResultCode UnmapPages(VAddr addr, KPageLinkedList& page_linked_list, KMemoryState state);
-    ResultCode SetProcessMemoryPermission(VAddr addr, std::size_t size, KMemoryPermission perm);
+    ResultCode SetProcessMemoryPermission(VAddr addr, std::size_t size,
+                                          Svc::MemoryPermission svc_perm);
     KMemoryInfo QueryInfo(VAddr addr);
     ResultCode ReserveTransferMemory(VAddr addr, std::size_t size, KMemoryPermission perm);
     ResultCode ResetTransferMemory(VAddr addr, std::size_t size);
     ResultCode SetMemoryPermission(VAddr addr, std::size_t size, Svc::MemoryPermission perm);
-    ResultCode SetMemoryAttribute(VAddr addr, std::size_t size, KMemoryAttribute mask,
-                                  KMemoryAttribute value);
+    ResultCode SetMemoryAttribute(VAddr addr, std::size_t size, u32 mask, u32 attr);
     ResultCode SetMaxHeapSize(std::size_t size);
     ResultCode SetHeapSize(VAddr* out, std::size_t size);
     ResultVal<VAddr> AllocateAndMapMemory(std::size_t needed_num_pages, std::size_t align,
@@ -88,7 +92,6 @@ private:
     ResultCode MapPages(VAddr addr, const KPageLinkedList& page_linked_list,
                         KMemoryPermission perm);
     ResultCode UnmapPages(VAddr addr, const KPageLinkedList& page_linked_list);
-    void MapPhysicalMemory(KPageLinkedList& page_linked_list, VAddr start, VAddr end);
     bool IsRegionMapped(VAddr address, u64 size);
     bool IsRegionContiguous(VAddr addr, u64 size) const;
     void AddRegionToPages(VAddr start, std::size_t num_pages, KPageLinkedList& page_linked_list);
@@ -102,30 +105,54 @@ private:
     constexpr VAddr GetRegionAddress(KMemoryState state) const;
     constexpr std::size_t GetRegionSize(KMemoryState state) const;
 
-    constexpr ResultCode CheckMemoryState(const KMemoryInfo& info, KMemoryState state_mask,
+    ResultCode CheckMemoryStateContiguous(std::size_t* out_blocks_needed, VAddr addr,
+                                          std::size_t size, KMemoryState state_mask,
                                           KMemoryState state, KMemoryPermission perm_mask,
                                           KMemoryPermission perm, KMemoryAttribute attr_mask,
                                           KMemoryAttribute attr) const;
+    ResultCode CheckMemoryStateContiguous(VAddr addr, std::size_t size, KMemoryState state_mask,
+                                          KMemoryState state, KMemoryPermission perm_mask,
+                                          KMemoryPermission perm, KMemoryAttribute attr_mask,
+                                          KMemoryAttribute attr) const {
+        return this->CheckMemoryStateContiguous(nullptr, addr, size, state_mask, state, perm_mask,
+                                                perm, attr_mask, attr);
+    }
+
+    ResultCode CheckMemoryState(const KMemoryInfo& info, KMemoryState state_mask,
+                                KMemoryState state, KMemoryPermission perm_mask,
+                                KMemoryPermission perm, KMemoryAttribute attr_mask,
+                                KMemoryAttribute attr) const;
     ResultCode CheckMemoryState(KMemoryState* out_state, KMemoryPermission* out_perm,
-                                KMemoryAttribute* out_attr, VAddr addr, std::size_t size,
-                                KMemoryState state_mask, KMemoryState state,
-                                KMemoryPermission perm_mask, KMemoryPermission perm,
-                                KMemoryAttribute attr_mask, KMemoryAttribute attr,
-                                KMemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr);
-    ResultCode CheckMemoryState(VAddr addr, std::size_t size, KMemoryState state_mask,
+                                KMemoryAttribute* out_attr, std::size_t* out_blocks_needed,
+                                VAddr addr, std::size_t size, KMemoryState state_mask,
                                 KMemoryState state, KMemoryPermission perm_mask,
                                 KMemoryPermission perm, KMemoryAttribute attr_mask,
                                 KMemoryAttribute attr,
-                                KMemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr) {
-        return CheckMemoryState(nullptr, nullptr, nullptr, addr, size, state_mask, state, perm_mask,
-                                perm, attr_mask, attr, ignore_attr);
-    }
-    ResultCode CheckMemoryState(size_t* out_blocks_needed, VAddr addr, size_t size,
+                                KMemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr) const;
+    ResultCode CheckMemoryState(std::size_t* out_blocks_needed, VAddr addr, std::size_t size,
                                 KMemoryState state_mask, KMemoryState state,
                                 KMemoryPermission perm_mask, KMemoryPermission perm,
-                                KMemoryAttribute attr_mask, KMemoryAttribute attr) const;
+                                KMemoryAttribute attr_mask, KMemoryAttribute attr,
+                                KMemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr) const {
+        return CheckMemoryState(nullptr, nullptr, nullptr, out_blocks_needed, addr, size,
+                                state_mask, state, perm_mask, perm, attr_mask, attr, ignore_attr);
+    }
+    ResultCode CheckMemoryState(VAddr addr, size_t size, KMemoryState state_mask,
+                                KMemoryState state, KMemoryPermission perm_mask,
+                                KMemoryPermission perm, KMemoryAttribute attr_mask,
+                                KMemoryAttribute attr,
+                                KMemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr) const {
+        return this->CheckMemoryState(nullptr, addr, size, state_mask, state, perm_mask, perm,
+                                      attr_mask, attr, ignore_attr);
+    }
+
+    bool IsLockedByCurrentThread() const {
+        return general_lock.IsLockedByCurrentThread();
+    }
+
+    mutable KLightLock general_lock;
+    mutable KLightLock map_physical_memory_lock;
 
-    std::recursive_mutex page_table_lock;
     std::unique_ptr<KMemoryBlockManager> block_manager;
 
 public:
@@ -184,7 +211,7 @@ public:
         return alias_code_region_end - alias_code_region_start;
     }
     size_t GetNormalMemorySize() {
-        std::lock_guard lk(page_table_lock);
+        KScopedLightLock lk(general_lock);
         return GetHeapSize() + mapped_physical_memory_size;
     }
     constexpr std::size_t GetAddressSpaceWidth() const {
@@ -226,8 +253,12 @@ public:
     constexpr bool IsInsideASLRRegion(VAddr address, std::size_t size) const {
         return !IsOutsideASLRRegion(address, size);
     }
-    constexpr PAddr GetPhysicalAddr(VAddr addr) {
-        return page_table_impl.backing_addr[addr >> PageBits] + addr;
+
+    PAddr GetPhysicalAddr(VAddr addr) {
+        ASSERT(IsLockedByCurrentThread());
+        const auto backing_addr = page_table_impl.backing_addr[addr >> PageBits];
+        ASSERT(backing_addr);
+        return backing_addr + addr;
     }
     constexpr bool Contains(VAddr addr) const {
         return address_space_start <= addr && addr <= address_space_end - 1;
@@ -281,6 +312,7 @@ private:
     bool is_aslr_enabled{};
 
     KMemoryManager::Pool memory_pool{KMemoryManager::Pool::Application};
+    KMemoryManager::Direction allocation_option{KMemoryManager::Direction::FromFront};
 
     Common::PageTable page_table_impl;
 

src/core/hle/kernel/k_priority_queue.h

@@ -45,6 +45,7 @@ concept KPriorityQueueMember = !std::is_reference_v<T> && requires(T & t) {
 
     { t.GetActiveCore() } -> Common::ConvertibleTo<s32>;
     { t.GetPriority() } -> Common::ConvertibleTo<s32>;
+    { t.IsDummyThread() } -> Common::ConvertibleTo<bool>;
 };
 
 template <typename Member, size_t NumCores_, int LowestPriority, int HighestPriority>
@@ -257,7 +258,7 @@ private:
 
 private:
     constexpr void ClearAffinityBit(u64& affinity, s32 core) {
-        affinity &= ~(u64(1) << core);
+        affinity &= ~(UINT64_C(1) << core);
     }
 
     constexpr s32 GetNextCore(u64& affinity) {
@@ -349,24 +350,49 @@ public:
 
     // Mutators.
     constexpr void PushBack(Member* member) {
+        // This is for host (dummy) threads that we do not want to enter the priority queue.
+        if (member->IsDummyThread()) {
+            return;
+        }
+
         this->PushBack(member->GetPriority(), member);
     }
 
     constexpr void Remove(Member* member) {
+        // This is for host (dummy) threads that we do not want to enter the priority queue.
+        if (member->IsDummyThread()) {
+            return;
+        }
+
         this->Remove(member->GetPriority(), member);
     }
 
     constexpr void MoveToScheduledFront(Member* member) {
+        // This is for host (dummy) threads that we do not want to enter the priority queue.
+        if (member->IsDummyThread()) {
+            return;
+        }
+
         this->scheduled_queue.MoveToFront(member->GetPriority(), member->GetActiveCore(), member);
     }
 
     constexpr KThread* MoveToScheduledBack(Member* member) {
+        // This is for host (dummy) threads that we do not want to enter the priority queue.
+        if (member->IsDummyThread()) {
+            return {};
+        }
+
         return this->scheduled_queue.MoveToBack(member->GetPriority(), member->GetActiveCore(),
                                                 member);
     }
 
     // First class fancy operations.
     constexpr void ChangePriority(s32 prev_priority, bool is_running, Member* member) {
+        // This is for host (dummy) threads that we do not want to enter the priority queue.
+        if (member->IsDummyThread()) {
+            return;
+        }
+
         ASSERT(IsValidPriority(prev_priority));
 
         // Remove the member from the queues.
@@ -383,6 +409,11 @@ public:
 
     constexpr void ChangeAffinityMask(s32 prev_core, const AffinityMaskType& prev_affinity,
                                       Member* member) {
+        // This is for host (dummy) threads that we do not want to enter the priority queue.
+        if (member->IsDummyThread()) {
+            return;
+        }
+
         // Get the new information.
         const s32 priority = member->GetPriority();
         const AffinityMaskType& new_affinity = member->GetAffinityMask();
@@ -412,6 +443,11 @@ public:
     }
 
     constexpr void ChangeCore(s32 prev_core, Member* member, bool to_front = false) {
+        // This is for host (dummy) threads that we do not want to enter the priority queue.
+        if (member->IsDummyThread()) {
+            return;
+        }
+
         // Get the new information.
         const s32 new_core = member->GetActiveCore();
         const s32 priority = member->GetPriority();

src/core/hle/kernel/k_process.cpp

@@ -146,27 +146,35 @@ ResultCode KProcess::Initialize(KProcess* process, Core::System& system, std::st
     // Open a reference to the resource limit.
     process->resource_limit->Open();
 
+    // Clear remaining fields.
+    process->num_running_threads = 0;
+    process->is_signaled = false;
+    process->exception_thread = nullptr;
+    process->is_suspended = false;
+    process->schedule_count = 0;
+
     return ResultSuccess;
 }
 
+void KProcess::DoWorkerTaskImpl() {
+    UNIMPLEMENTED();
+}
+
 KResourceLimit* KProcess::GetResourceLimit() const {
     return resource_limit;
 }
 
-void KProcess::IncrementThreadCount() {
-    ASSERT(num_threads >= 0);
-    num_created_threads++;
-
-    if (const auto count = ++num_threads; count > peak_num_threads) {
-        peak_num_threads = count;
-    }
+void KProcess::IncrementRunningThreadCount() {
+    ASSERT(num_running_threads.load() >= 0);
+    ++num_running_threads;
 }
 
-void KProcess::DecrementThreadCount() {
-    ASSERT(num_threads > 0);
+void KProcess::DecrementRunningThreadCount() {
+    ASSERT(num_running_threads.load() > 0);
 
-    if (const auto count = --num_threads; count == 0) {
-        LOG_WARNING(Kernel, "Process termination is not fully implemented.");
+    if (const auto prev = num_running_threads--; prev == 1) {
+        // TODO(bunnei): Process termination to be implemented when multiprocess is supported.
+        UNIMPLEMENTED_MSG("KProcess termination is not implemennted!");
     }
 }
 
@@ -220,30 +228,28 @@ bool KProcess::ReleaseUserException(KThread* thread) {
     }
 }
 
-void KProcess::PinCurrentThread() {
+void KProcess::PinCurrentThread(s32 core_id) {
     ASSERT(kernel.GlobalSchedulerContext().IsLocked());
 
     // Get the current thread.
-    const s32 core_id = GetCurrentCoreId(kernel);
-    KThread* cur_thread = GetCurrentThreadPointer(kernel);
+    KThread* cur_thread = kernel.Scheduler(static_cast<std::size_t>(core_id)).GetCurrentThread();
 
     // If the thread isn't terminated, pin it.
     if (!cur_thread->IsTerminationRequested()) {
         // Pin it.
         PinThread(core_id, cur_thread);
-        cur_thread->Pin();
+        cur_thread->Pin(core_id);
 
         // An update is needed.
         KScheduler::SetSchedulerUpdateNeeded(kernel);
     }
 }
 
-void KProcess::UnpinCurrentThread() {
+void KProcess::UnpinCurrentThread(s32 core_id) {
     ASSERT(kernel.GlobalSchedulerContext().IsLocked());
 
     // Get the current thread.
-    const s32 core_id = GetCurrentCoreId(kernel);
-    KThread* cur_thread = GetCurrentThreadPointer(kernel);
+    KThread* cur_thread = kernel.Scheduler(static_cast<std::size_t>(core_id)).GetCurrentThread();
 
     // Unpin it.
     cur_thread->Unpin();
@@ -479,7 +485,7 @@ void KProcess::Finalize() {
     }
 
     // Perform inherited finalization.
-    KAutoObjectWithSlabHeapAndContainer<KProcess, KSynchronizationObject>::Finalize();
+    KAutoObjectWithSlabHeapAndContainer<KProcess, KWorkerTask>::Finalize();
 }
 
 /**
@@ -511,7 +517,7 @@ VAddr KProcess::CreateTLSRegion() {
     const VAddr tls_page_addr{page_table
                                   ->AllocateAndMapMemory(1, PageSize, true, start, size / PageSize,
                                                          KMemoryState::ThreadLocal,
-                                                         KMemoryPermission::ReadAndWrite,
+                                                         KMemoryPermission::UserReadWrite,
                                                          tls_map_addr)
                                   .ValueOr(0)};
 
@@ -543,16 +549,16 @@ void KProcess::FreeTLSRegion(VAddr tls_address) {
 
 void KProcess::LoadModule(CodeSet code_set, VAddr base_addr) {
     const auto ReprotectSegment = [&](const CodeSet::Segment& segment,
-                                      KMemoryPermission permission) {
+                                      Svc::MemoryPermission permission) {
         page_table->SetProcessMemoryPermission(segment.addr + base_addr, segment.size, permission);
     };
 
     kernel.System().Memory().WriteBlock(*this, base_addr, code_set.memory.data(),
                                         code_set.memory.size());
 
-    ReprotectSegment(code_set.CodeSegment(), KMemoryPermission::ReadAndExecute);
-    ReprotectSegment(code_set.RODataSegment(), KMemoryPermission::Read);
-    ReprotectSegment(code_set.DataSegment(), KMemoryPermission::ReadAndWrite);
+    ReprotectSegment(code_set.CodeSegment(), Svc::MemoryPermission::ReadExecute);
+    ReprotectSegment(code_set.RODataSegment(), Svc::MemoryPermission::Read);
+    ReprotectSegment(code_set.DataSegment(), Svc::MemoryPermission::ReadWrite);
 }
 
 bool KProcess::IsSignaled() const {
@@ -589,7 +595,7 @@ ResultCode KProcess::AllocateMainThreadStack(std::size_t stack_size) {
     CASCADE_RESULT(main_thread_stack_top,
                    page_table->AllocateAndMapMemory(
                        main_thread_stack_size / PageSize, PageSize, false, start, size / PageSize,
-                       KMemoryState::Stack, KMemoryPermission::ReadAndWrite));
+                       KMemoryState::Stack, KMemoryPermission::UserReadWrite));
 
     main_thread_stack_top += main_thread_stack_size;
 

src/core/hle/kernel/k_process.h

@@ -15,6 +15,7 @@
 #include "core/hle/kernel/k_condition_variable.h"
 #include "core/hle/kernel/k_handle_table.h"
 #include "core/hle/kernel/k_synchronization_object.h"
+#include "core/hle/kernel/k_worker_task.h"
 #include "core/hle/kernel/process_capability.h"
 #include "core/hle/kernel/slab_helpers.h"
 #include "core/hle/result.h"
@@ -62,8 +63,7 @@ enum class ProcessStatus {
     DebugBreak,
 };
 
-class KProcess final
-    : public KAutoObjectWithSlabHeapAndContainer<KProcess, KSynchronizationObject> {
+class KProcess final : public KAutoObjectWithSlabHeapAndContainer<KProcess, KWorkerTask> {
     KERNEL_AUTOOBJECT_TRAITS(KProcess, KSynchronizationObject);
 
 public:
@@ -235,8 +235,8 @@ public:
         ++schedule_count;
     }
 
-    void IncrementThreadCount();
-    void DecrementThreadCount();
+    void IncrementRunningThreadCount();
+    void DecrementRunningThreadCount();
 
     void SetRunningThread(s32 core, KThread* thread, u64 idle_count) {
         running_threads[core] = thread;
@@ -345,8 +345,10 @@ public:
 
     bool IsSignaled() const override;
 
-    void PinCurrentThread();
-    void UnpinCurrentThread();
+    void DoWorkerTaskImpl();
+
+    void PinCurrentThread(s32 core_id);
+    void UnpinCurrentThread(s32 core_id);
     void UnpinThread(KThread* thread);
 
     KLightLock& GetStateLock() {
@@ -471,9 +473,7 @@ private:
     bool is_suspended{};
     bool is_initialized{};
 
-    std::atomic<s32> num_created_threads{};
-    std::atomic<u16> num_threads{};
-    u16 peak_num_threads{};
+    std::atomic<u16> num_running_threads{};
 
     std::array<KThread*, Core::Hardware::NUM_CPU_CORES> running_threads{};
     std::array<u64, Core::Hardware::NUM_CPU_CORES> running_thread_idle_counts{};

src/core/hle/kernel/k_scheduler.cpp

@@ -15,6 +15,7 @@
 #include "core/core.h"
 #include "core/core_timing.h"
 #include "core/cpu_manager.h"
+#include "core/hle/kernel/k_interrupt_manager.h"
 #include "core/hle/kernel/k_process.h"
 #include "core/hle/kernel/k_scheduler.h"
 #include "core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h"
@@ -48,11 +49,16 @@ void KScheduler::RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedul
         if (!must_context_switch || core != current_core) {
             auto& phys_core = kernel.PhysicalCore(core);
             phys_core.Interrupt();
-        } else {
-            must_context_switch = true;
         }
         cores_pending_reschedule &= ~(1ULL << core);
     }
+
+    for (std::size_t core_id = 0; core_id < Core::Hardware::NUM_CPU_CORES; ++core_id) {
+        if (kernel.PhysicalCore(core_id).IsInterrupted()) {
+            KInterruptManager::HandleInterrupt(kernel, static_cast<s32>(core_id));
+        }
+    }
+
     if (must_context_switch) {
         auto core_scheduler = kernel.CurrentScheduler();
         kernel.ExitSVCProfile();
@@ -400,6 +406,9 @@ void KScheduler::EnableScheduling(KernelCore& kernel, u64 cores_needing_scheduli
     } else {
         RescheduleCores(kernel, cores_needing_scheduling);
     }
+
+    // Special case to ensure dummy threads that are waiting block.
+    current_thread->IfDummyThreadTryWait();
 }
 
 u64 KScheduler::UpdateHighestPriorityThreads(KernelCore& kernel) {
@@ -701,23 +710,19 @@ void KScheduler::Unload(KThread* thread) {
 }
 
 void KScheduler::Reload(KThread* thread) {
-    LOG_TRACE(Kernel, "core {}, reload thread {}", core_id, thread ? thread->GetName() : "nullptr");
+    LOG_TRACE(Kernel, "core {}, reload thread {}", core_id, thread->GetName());
 
-    if (thread) {
-        ASSERT_MSG(thread->GetState() == ThreadState::Runnable, "Thread must be runnable.");
-
-        Core::ARM_Interface& cpu_core = system.ArmInterface(core_id);
-        cpu_core.LoadContext(thread->GetContext32());
-        cpu_core.LoadContext(thread->GetContext64());
-        cpu_core.SetTlsAddress(thread->GetTLSAddress());
-        cpu_core.SetTPIDR_EL0(thread->GetTPIDR_EL0());
-        cpu_core.ClearExclusiveState();
-    }
+    Core::ARM_Interface& cpu_core = system.ArmInterface(core_id);
+    cpu_core.LoadContext(thread->GetContext32());
+    cpu_core.LoadContext(thread->GetContext64());
+    cpu_core.SetTlsAddress(thread->GetTLSAddress());
+    cpu_core.SetTPIDR_EL0(thread->GetTPIDR_EL0());
+    cpu_core.ClearExclusiveState();
 }
 
 void KScheduler::SwitchContextStep2() {
     // Load context of new thread
-    Reload(current_thread.load());
+    Reload(GetCurrentThread());
 
     RescheduleCurrentCore();
 }
@@ -726,13 +731,23 @@ void KScheduler::ScheduleImpl() {
     KThread* previous_thread = GetCurrentThread();
     KThread* next_thread = state.highest_priority_thread;
 
-    state.needs_scheduling = false;
+    state.needs_scheduling.store(false);
 
     // We never want to schedule a null thread, so use the idle thread if we don't have a next.
     if (next_thread == nullptr) {
         next_thread = idle_thread;
     }
 
+    if (next_thread->GetCurrentCore() != core_id) {
+        next_thread->SetCurrentCore(core_id);
+    }
+
+    // We never want to schedule a dummy thread, as these are only used by host threads for locking.
+    if (next_thread->GetThreadType() == ThreadType::Dummy) {
+        ASSERT_MSG(false, "Dummy threads should never be scheduled!");
+        next_thread = idle_thread;
+    }
+
     // If we're not actually switching thread, there's nothing to do.
     if (next_thread == current_thread.load()) {
         previous_thread->EnableDispatch();
@@ -740,14 +755,8 @@ void KScheduler::ScheduleImpl() {
         return;
     }
 
-    if (next_thread->GetCurrentCore() != core_id) {
-        next_thread->SetCurrentCore(core_id);
-    }
-
-    current_thread.store(next_thread);
-
+    // Update the CPU time tracking variables.
     KProcess* const previous_process = system.Kernel().CurrentProcess();
-
     UpdateLastContextSwitchTime(previous_thread, previous_process);
 
     // Save context for previous thread
@@ -755,6 +764,10 @@ void KScheduler::ScheduleImpl() {
 
     std::shared_ptr<Common::Fiber>* old_context;
     old_context = &previous_thread->GetHostContext();
+
+    // Set the new thread.
+    current_thread.store(next_thread);
+
     guard.Unlock();
 
     Common::Fiber::YieldTo(*old_context, *switch_fiber);
@@ -782,8 +795,8 @@ void KScheduler::SwitchToCurrent() {
         do {
             auto next_thread = current_thread.load();
             if (next_thread != nullptr) {
-                next_thread->context_guard.Lock();
-                if (next_thread->GetRawState() != ThreadState::Runnable) {
+                const auto locked = next_thread->context_guard.TryLock();
+                if (state.needs_scheduling.load()) {
                     next_thread->context_guard.Unlock();
                     break;
                 }
@@ -791,6 +804,9 @@ void KScheduler::SwitchToCurrent() {
                     next_thread->context_guard.Unlock();
                     break;
                 }
+                if (!locked) {
+                    continue;
+                }
             }
             auto thread = next_thread ? next_thread : idle_thread;
             Common::Fiber::YieldTo(switch_fiber, *thread->GetHostContext());

src/core/hle/kernel/k_server_session.cpp

@@ -8,7 +8,6 @@
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "common/logging/log.h"
-#include "common/scope_exit.h"
 #include "core/core_timing.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/hle_ipc.h"
@@ -123,20 +122,10 @@ ResultCode KServerSession::QueueSyncRequest(KThread* thread, Core::Memory::Memor
 
     context->PopulateFromIncomingCommandBuffer(kernel.CurrentProcess()->GetHandleTable(), cmd_buf);
 
-    // In the event that something fails here, stub a result to prevent the game from crashing.
-    // This is a work-around in the event that somehow we process a service request after the
-    // session has been closed by the game. This has been observed to happen rarely in Pokemon
-    // Sword/Shield and is likely a result of us using host threads/scheduling for services.
-    // TODO(bunnei): Find a better solution here.
-    auto error_guard = SCOPE_GUARD({ CompleteSyncRequest(*context); });
-
     // Ensure we have a session request handler
     if (manager->HasSessionRequestHandler(*context)) {
         if (auto strong_ptr = manager->GetServiceThread().lock()) {
             strong_ptr->QueueSyncRequest(*parent, std::move(context));
-
-            // We succeeded.
-            error_guard.Cancel();
         } else {
             ASSERT_MSG(false, "strong_ptr is nullptr!");
         }
@@ -171,13 +160,8 @@ ResultCode KServerSession::CompleteSyncRequest(HLERequestContext& context) {
         convert_to_domain = false;
     }
 
-    // Some service requests require the thread to block
-    {
-        KScopedSchedulerLock lock(kernel);
-        if (!context.IsThreadWaiting()) {
-            context.GetThread().EndWait(result);
-        }
-    }
+    // The calling thread is waiting for this request to complete, so wake it up.
+    context.GetThread().EndWait(result);
 
     return result;
 }

src/core/hle/kernel/k_slab_heap.h

@@ -7,6 +7,7 @@
 #include <atomic>
 
 #include "common/assert.h"
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 
 namespace Kernel {
@@ -15,13 +16,17 @@ class KernelCore;
 
 namespace impl {
 
-class KSlabHeapImpl final : NonCopyable {
+class KSlabHeapImpl final {
 public:
+    YUZU_NON_COPYABLE(KSlabHeapImpl);
+    YUZU_NON_MOVEABLE(KSlabHeapImpl);
+
     struct Node {
         Node* next{};
     };
 
     constexpr KSlabHeapImpl() = default;
+    constexpr ~KSlabHeapImpl() = default;
 
     void Initialize(std::size_t size) {
         ASSERT(head == nullptr);
@@ -64,9 +69,13 @@ private:
 
 } // namespace impl
 
-class KSlabHeapBase : NonCopyable {
+class KSlabHeapBase {
 public:
+    YUZU_NON_COPYABLE(KSlabHeapBase);
+    YUZU_NON_MOVEABLE(KSlabHeapBase);
+
     constexpr KSlabHeapBase() = default;
+    constexpr ~KSlabHeapBase() = default;
 
     constexpr bool Contains(uintptr_t addr) const {
         return start <= addr && addr < end;

src/core/hle/kernel/k_thread.cpp

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <algorithm>
+#include <atomic>
 #include <cinttypes>
 #include <optional>
 #include <vector>
@@ -29,10 +30,12 @@
 #include "core/hle/kernel/k_system_control.h"
 #include "core/hle/kernel/k_thread.h"
 #include "core/hle/kernel/k_thread_queue.h"
+#include "core/hle/kernel/k_worker_task_manager.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/svc_results.h"
 #include "core/hle/kernel/time_manager.h"
 #include "core/hle/result.h"
+#include "core/memory.h"
 
 #ifdef ARCHITECTURE_x86_64
 #include "core/arm/dynarmic/arm_dynarmic_32.h"
@@ -63,6 +66,13 @@ namespace Kernel {
 
 namespace {
 
+struct ThreadLocalRegion {
+    static constexpr std::size_t MessageBufferSize = 0x100;
+    std::array<u32, MessageBufferSize / sizeof(u32)> message_buffer;
+    std::atomic_uint16_t disable_count;
+    std::atomic_uint16_t interrupt_flag;
+};
+
 class ThreadQueueImplForKThreadSleep final : public KThreadQueueWithoutEndWait {
 public:
     explicit ThreadQueueImplForKThreadSleep(KernelCore& kernel_)
@@ -96,7 +106,7 @@ KThread::~KThread() = default;
 ResultCode KThread::Initialize(KThreadFunction func, uintptr_t arg, VAddr user_stack_top, s32 prio,
                                s32 virt_core, KProcess* owner, ThreadType type) {
     // Assert parameters are valid.
-    ASSERT((type == ThreadType::Main) ||
+    ASSERT((type == ThreadType::Main) || (type == ThreadType::Dummy) ||
            (Svc::HighestThreadPriority <= prio && prio <= Svc::LowestThreadPriority));
     ASSERT((owner != nullptr) || (type != ThreadType::User));
     ASSERT(0 <= virt_core && virt_core < static_cast<s32>(Common::BitSize<u64>()));
@@ -130,7 +140,7 @@ ResultCode KThread::Initialize(KThreadFunction func, uintptr_t arg, VAddr user_s
         UNREACHABLE_MSG("KThread::Initialize: Unknown ThreadType {}", static_cast<u32>(type));
         break;
     }
-    thread_type_for_debugging = type;
+    thread_type = type;
 
     // Set the ideal core ID and affinity mask.
     virtual_ideal_core_id = virt_core;
@@ -205,7 +215,6 @@ ResultCode KThread::Initialize(KThreadFunction func, uintptr_t arg, VAddr user_s
 
         parent = owner;
         parent->Open();
-        parent->IncrementThreadCount();
     }
 
     // Initialize thread context.
@@ -252,7 +261,7 @@ ResultCode KThread::InitializeThread(KThread* thread, KThreadFunction func, uint
 }
 
 ResultCode KThread::InitializeDummyThread(KThread* thread) {
-    return thread->Initialize({}, {}, {}, DefaultThreadPriority, 3, {}, ThreadType::Dummy);
+    return thread->Initialize({}, {}, {}, DummyThreadPriority, 3, {}, ThreadType::Dummy);
 }
 
 ResultCode KThread::InitializeIdleThread(Core::System& system, KThread* thread, s32 virt_core) {
@@ -317,13 +326,8 @@ void KThread::Finalize() {
         }
     }
 
-    // Decrement the parent process's thread count.
-    if (parent != nullptr) {
-        parent->DecrementThreadCount();
-    }
-
     // Perform inherited finalization.
-    KAutoObjectWithSlabHeapAndContainer<KThread, KSynchronizationObject>::Finalize();
+    KSynchronizationObject::Finalize();
 }
 
 bool KThread::IsSignaled() const {
@@ -346,7 +350,7 @@ void KThread::StartTermination() {
     if (parent != nullptr) {
         parent->ReleaseUserException(this);
         if (parent->GetPinnedThread(GetCurrentCoreId(kernel)) == this) {
-            parent->UnpinCurrentThread();
+            parent->UnpinCurrentThread(core_id);
         }
     }
 
@@ -367,12 +371,29 @@ void KThread::StartTermination() {
 
     // Register terminated dpc flag.
     RegisterDpc(DpcFlag::Terminated);
+}
+
+void KThread::FinishTermination() {
+    // Ensure that the thread is not executing on any core.
+    if (parent != nullptr) {
+        for (std::size_t i = 0; i < static_cast<std::size_t>(Core::Hardware::NUM_CPU_CORES); ++i) {
+            KThread* core_thread{};
+            do {
+                core_thread = kernel.Scheduler(i).GetCurrentThread();
+            } while (core_thread == this);
+        }
+    }
 
     // Close the thread.
     this->Close();
 }
 
-void KThread::Pin() {
+void KThread::DoWorkerTaskImpl() {
+    // Finish the termination that was begun by Exit().
+    this->FinishTermination();
+}
+
+void KThread::Pin(s32 current_core) {
     ASSERT(kernel.GlobalSchedulerContext().IsLocked());
 
     // Set ourselves as pinned.
@@ -389,7 +410,6 @@ void KThread::Pin() {
 
         // Bind ourselves to this core.
         const s32 active_core = GetActiveCore();
-        const s32 current_core = GetCurrentCoreId(kernel);
 
         SetActiveCore(current_core);
         physical_ideal_core_id = current_core;
@@ -409,12 +429,7 @@ void KThread::Pin() {
                                          static_cast<u32>(ThreadState::SuspendShift)));
 
         // Update our state.
-        const ThreadState old_state = thread_state;
-        thread_state = static_cast<ThreadState>(GetSuspendFlags() |
-                                                static_cast<u32>(old_state & ThreadState::Mask));
-        if (thread_state != old_state) {
-            KScheduler::OnThreadStateChanged(kernel, this, old_state);
-        }
+        UpdateState();
     }
 
     // TODO(bunnei): Update our SVC access permissions.
@@ -455,20 +470,13 @@ void KThread::Unpin() {
     }
 
     // Allow performing thread suspension (if termination hasn't been requested).
-    {
+    if (!IsTerminationRequested()) {
         // Update our allow flags.
-        if (!IsTerminationRequested()) {
-            suspend_allowed_flags |= (1 << (static_cast<u32>(SuspendType::Thread) +
-                                            static_cast<u32>(ThreadState::SuspendShift)));
-        }
+        suspend_allowed_flags |= (1 << (static_cast<u32>(SuspendType::Thread) +
+                                        static_cast<u32>(ThreadState::SuspendShift)));
 
         // Update our state.
-        const ThreadState old_state = thread_state;
-        thread_state = static_cast<ThreadState>(GetSuspendFlags() |
-                                                static_cast<u32>(old_state & ThreadState::Mask));
-        if (thread_state != old_state) {
-            KScheduler::OnThreadStateChanged(kernel, this, old_state);
-        }
+        UpdateState();
     }
 
     // TODO(bunnei): Update our SVC access permissions.
@@ -482,6 +490,36 @@ void KThread::Unpin() {
     }
 }
 
+u16 KThread::GetUserDisableCount() const {
+    if (!IsUserThread()) {
+        // We only emulate TLS for user threads
+        return {};
+    }
+
+    auto& memory = kernel.System().Memory();
+    return memory.Read16(tls_address + offsetof(ThreadLocalRegion, disable_count));
+}
+
+void KThread::SetInterruptFlag() {
+    if (!IsUserThread()) {
+        // We only emulate TLS for user threads
+        return;
+    }
+
+    auto& memory = kernel.System().Memory();
+    memory.Write16(tls_address + offsetof(ThreadLocalRegion, interrupt_flag), 1);
+}
+
+void KThread::ClearInterruptFlag() {
+    if (!IsUserThread()) {
+        // We only emulate TLS for user threads
+        return;
+    }
+
+    auto& memory = kernel.System().Memory();
+    memory.Write16(tls_address + offsetof(ThreadLocalRegion, interrupt_flag), 0);
+}
+
 ResultCode KThread::GetCoreMask(s32* out_ideal_core, u64* out_affinity_mask) {
     KScopedSchedulerLock sl{kernel};
 
@@ -651,12 +689,7 @@ void KThread::Resume(SuspendType type) {
         ~(1u << (static_cast<u32>(ThreadState::SuspendShift) + static_cast<u32>(type)));
 
     // Update our state.
-    const ThreadState old_state = thread_state;
-    thread_state = static_cast<ThreadState>(GetSuspendFlags() |
-                                            static_cast<u32>(old_state & ThreadState::Mask));
-    if (thread_state != old_state) {
-        KScheduler::OnThreadStateChanged(kernel, this, old_state);
-    }
+    this->UpdateState();
 }
 
 void KThread::WaitCancel() {
@@ -683,19 +716,22 @@ void KThread::TrySuspend() {
     ASSERT(GetNumKernelWaiters() == 0);
 
     // Perform the suspend.
-    Suspend();
+    this->UpdateState();
 }
 
-void KThread::Suspend() {
+void KThread::UpdateState() {
     ASSERT(kernel.GlobalSchedulerContext().IsLocked());
-    ASSERT(IsSuspendRequested());
 
     // Set our suspend flags in state.
     const auto old_state = thread_state;
-    thread_state = static_cast<ThreadState>(GetSuspendFlags()) | (old_state & ThreadState::Mask);
+    const auto new_state =
+        static_cast<ThreadState>(this->GetSuspendFlags()) | (old_state & ThreadState::Mask);
+    thread_state = new_state;
 
     // Note the state change in scheduler.
-    KScheduler::OnThreadStateChanged(kernel, this, old_state);
+    if (new_state != old_state) {
+        KScheduler::OnThreadStateChanged(kernel, this, old_state);
+    }
 }
 
 void KThread::Continue() {
@@ -960,13 +996,16 @@ ResultCode KThread::Run() {
 
         // If the current thread has been asked to suspend, suspend it and retry.
         if (GetCurrentThread(kernel).IsSuspended()) {
-            GetCurrentThread(kernel).Suspend();
+            GetCurrentThread(kernel).UpdateState();
             continue;
         }
 
         // If we're not a kernel thread and we've been asked to suspend, suspend ourselves.
-        if (IsUserThread() && IsSuspended()) {
-            Suspend();
+        if (KProcess* owner = this->GetOwnerProcess(); owner != nullptr) {
+            if (IsUserThread() && IsSuspended()) {
+                this->UpdateState();
+            }
+            owner->IncrementRunningThreadCount();
         }
 
         // Set our state and finish.
@@ -981,21 +1020,29 @@ ResultCode KThread::Run() {
 void KThread::Exit() {
     ASSERT(this == GetCurrentThreadPointer(kernel));
 
-    // Release the thread resource hint from parent.
+    // Release the thread resource hint, running thread count from parent.
     if (parent != nullptr) {
         parent->GetResourceLimit()->Release(Kernel::LimitableResource::Threads, 0, 1);
         resource_limit_release_hint = true;
+        parent->DecrementRunningThreadCount();
     }
 
     // Perform termination.
     {
         KScopedSchedulerLock sl{kernel};
 
+        // Disallow all suspension.
+        suspend_allowed_flags = 0;
+        this->UpdateState();
+
         // Disallow all suspension.
         suspend_allowed_flags = 0;
 
         // Start termination.
         StartTermination();
+
+        // Register the thread as a work task.
+        KWorkerTaskManager::AddTask(kernel, KWorkerTaskManager::WorkerType::Exit, this);
     }
 }
 
@@ -1023,12 +1070,46 @@ ResultCode KThread::Sleep(s64 timeout) {
     return ResultSuccess;
 }
 
+void KThread::IfDummyThreadTryWait() {
+    if (!IsDummyThread()) {
+        return;
+    }
+
+    if (GetState() != ThreadState::Waiting) {
+        return;
+    }
+
+    // Block until we can grab the lock.
+    KScopedSpinLock lk{dummy_wait_lock};
+}
+
+void KThread::IfDummyThreadBeginWait() {
+    if (!IsDummyThread()) {
+        return;
+    }
+
+    // Ensure the thread will block when IfDummyThreadTryWait is called.
+    dummy_wait_lock.Lock();
+}
+
+void KThread::IfDummyThreadEndWait() {
+    if (!IsDummyThread()) {
+        return;
+    }
+
+    // Ensure the thread will no longer block.
+    dummy_wait_lock.Unlock();
+}
+
 void KThread::BeginWait(KThreadQueue* queue) {
     // Set our state as waiting.
     SetState(ThreadState::Waiting);
 
     // Set our wait queue.
     wait_queue = queue;
+
+    // Special case for dummy threads to ensure they block.
+    IfDummyThreadBeginWait();
 }
 
 void KThread::NotifyAvailable(KSynchronizationObject* signaled_object, ResultCode wait_result_) {
@@ -1047,7 +1128,16 @@ void KThread::EndWait(ResultCode wait_result_) {
 
     // If we're waiting, notify our queue that we're available.
     if (GetState() == ThreadState::Waiting) {
+        if (wait_queue == nullptr) {
+            // This should never happen, but avoid a hard crash below to get this logged.
+            ASSERT_MSG(false, "wait_queue is nullptr!");
+            return;
+        }
+
         wait_queue->EndWait(this, wait_result_);
+
+        // Special case for dummy threads to wakeup if necessary.
+        IfDummyThreadEndWait();
     }
 }
 

src/core/hle/kernel/k_thread.h

@@ -19,6 +19,7 @@
 #include "core/hle/kernel/k_light_lock.h"
 #include "core/hle/kernel/k_spin_lock.h"
 #include "core/hle/kernel/k_synchronization_object.h"
+#include "core/hle/kernel/k_worker_task.h"
 #include "core/hle/kernel/slab_helpers.h"
 #include "core/hle/kernel/svc_common.h"
 #include "core/hle/kernel/svc_types.h"
@@ -100,7 +101,7 @@ enum class ThreadWaitReasonForDebugging : u32 {
 [[nodiscard]] KThread& GetCurrentThread(KernelCore& kernel);
 [[nodiscard]] s32 GetCurrentCoreId(KernelCore& kernel);
 
-class KThread final : public KAutoObjectWithSlabHeapAndContainer<KThread, KSynchronizationObject>,
+class KThread final : public KAutoObjectWithSlabHeapAndContainer<KThread, KWorkerTask>,
                       public boost::intrusive::list_base_hook<> {
     KERNEL_AUTOOBJECT_TRAITS(KThread, KSynchronizationObject);
 
@@ -111,6 +112,7 @@ private:
 public:
     static constexpr s32 DefaultThreadPriority = 44;
     static constexpr s32 IdleThreadPriority = Svc::LowestThreadPriority + 1;
+    static constexpr s32 DummyThreadPriority = Svc::LowestThreadPriority + 2;
 
     explicit KThread(KernelCore& kernel_);
     ~KThread() override;
@@ -192,9 +194,9 @@ public:
 
     void TrySuspend();
 
-    void Continue();
+    void UpdateState();
 
-    void Suspend();
+    void Continue();
 
     constexpr void SetSyncedIndex(s32 index) {
         synced_index = index;
@@ -307,6 +309,10 @@ public:
         return parent != nullptr;
     }
 
+    u16 GetUserDisableCount() const;
+    void SetInterruptFlag();
+    void ClearInterruptFlag();
+
     [[nodiscard]] KThread* GetLockOwner() const {
         return lock_owner;
     }
@@ -381,6 +387,8 @@ public:
 
     void OnTimer();
 
+    void DoWorkerTaskImpl();
+
     static void PostDestroy(uintptr_t arg);
 
     [[nodiscard]] static ResultCode InitializeDummyThread(KThread* thread);
@@ -490,7 +498,7 @@ public:
         this->GetStackParameters().disable_count--;
     }
 
-    void Pin();
+    void Pin(s32 current_core);
 
     void Unpin();
 
@@ -546,8 +554,12 @@ public:
         return wait_reason_for_debugging;
     }
 
-    [[nodiscard]] ThreadType GetThreadTypeForDebugging() const {
-        return thread_type_for_debugging;
+    [[nodiscard]] ThreadType GetThreadType() const {
+        return thread_type;
+    }
+
+    [[nodiscard]] bool IsDummyThread() const {
+        return GetThreadType() == ThreadType::Dummy;
     }
 
     void SetWaitObjectsForDebugging(const std::span<KSynchronizationObject*>& objects) {
@@ -624,6 +636,14 @@ public:
         return condvar_key;
     }
 
+    // Dummy threads (used for HLE host threads) cannot wait based on the guest scheduler, and
+    // therefore will not block on guest kernel synchronization primitives. These methods handle
+    // blocking as needed.
+
+    void IfDummyThreadTryWait();
+    void IfDummyThreadBeginWait();
+    void IfDummyThreadEndWait();
+
 private:
     static constexpr size_t PriorityInheritanceCountMax = 10;
     union SyncObjectBuffer {
@@ -675,6 +695,8 @@ private:
 
     void StartTermination();
 
+    void FinishTermination();
+
     [[nodiscard]] ResultCode Initialize(KThreadFunction func, uintptr_t arg, VAddr user_stack_top,
                                         s32 prio, s32 virt_core, KProcess* owner, ThreadType type);
 
@@ -740,16 +762,17 @@ private:
     bool resource_limit_release_hint{};
     StackParameters stack_parameters{};
     KSpinLock context_guard{};
+    KSpinLock dummy_wait_lock{};
 
     // For emulation
     std::shared_ptr<Common::Fiber> host_context{};
     bool is_single_core{};
+    ThreadType thread_type{};
 
     // For debugging
     std::vector<KSynchronizationObject*> wait_objects_for_debugging;
     VAddr mutex_wait_address_for_debugging{};
     ThreadWaitReasonForDebugging wait_reason_for_debugging{};
-    ThreadType thread_type_for_debugging{};
 
 public:
     using ConditionVariableThreadTreeType = ConditionVariableThreadTree;

src/core/hle/kernel/k_worker_task.h

@@ -0,0 +1,18 @@
+// Copyright 2022 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/kernel/k_synchronization_object.h"
+
+namespace Kernel {
+
+class KWorkerTask : public KSynchronizationObject {
+public:
+    explicit KWorkerTask(KernelCore& kernel_);
+
+    void DoWorkerTask();
+};
+
+} // namespace Kernel

src/core/hle/kernel/k_worker_task_manager.cpp

@@ -0,0 +1,42 @@
+// Copyright 2022 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/assert.h"
+#include "core/hle/kernel/k_process.h"
+#include "core/hle/kernel/k_thread.h"
+#include "core/hle/kernel/k_worker_task.h"
+#include "core/hle/kernel/k_worker_task_manager.h"
+#include "core/hle/kernel/kernel.h"
+
+namespace Kernel {
+
+KWorkerTask::KWorkerTask(KernelCore& kernel_) : KSynchronizationObject{kernel_} {}
+
+void KWorkerTask::DoWorkerTask() {
+    if (auto* const thread = this->DynamicCast<KThread*>(); thread != nullptr) {
+        return thread->DoWorkerTaskImpl();
+    } else {
+        auto* const process = this->DynamicCast<KProcess*>();
+        ASSERT(process != nullptr);
+
+        return process->DoWorkerTaskImpl();
+    }
+}
+
+KWorkerTaskManager::KWorkerTaskManager() : m_waiting_thread(1, "yuzu:KWorkerTaskManager") {}
+
+void KWorkerTaskManager::AddTask(KernelCore& kernel, WorkerType type, KWorkerTask* task) {
+    ASSERT(type <= WorkerType::Count);
+    kernel.WorkerTaskManager().AddTask(kernel, task);
+}
+
+void KWorkerTaskManager::AddTask(KernelCore& kernel, KWorkerTask* task) {
+    KScopedSchedulerLock sl(kernel);
+    m_waiting_thread.QueueWork([task]() {
+        // Do the task.
+        task->DoWorkerTask();
+    });
+}
+
+} // namespace Kernel

src/core/hle/kernel/k_worker_task_manager.h

@@ -0,0 +1,33 @@
+// Copyright 2022 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+#include "common/thread_worker.h"
+
+namespace Kernel {
+
+class KernelCore;
+class KWorkerTask;
+
+class KWorkerTaskManager final {
+public:
+    enum class WorkerType : u32 {
+        Exit,
+        Count,
+    };
+
+    KWorkerTaskManager();
+
+    static void AddTask(KernelCore& kernel_, WorkerType type, KWorkerTask* task);
+
+private:
+    void AddTask(KernelCore& kernel, KWorkerTask* task);
+
+private:
+    Common::ThreadWorker m_waiting_thread;
+};
+
+} // namespace Kernel

src/core/hle/kernel/kernel.cpp

@@ -37,6 +37,7 @@
 #include "core/hle/kernel/k_shared_memory.h"
 #include "core/hle/kernel/k_slab_heap.h"
 #include "core/hle/kernel/k_thread.h"
+#include "core/hle/kernel/k_worker_task_manager.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/physical_core.h"
 #include "core/hle/kernel/service_thread.h"
@@ -51,7 +52,8 @@ namespace Kernel {
 
 struct KernelCore::Impl {
     explicit Impl(Core::System& system_, KernelCore& kernel_)
-        : time_manager{system_}, object_list_container{kernel_}, system{system_} {}
+        : time_manager{system_}, object_list_container{kernel_},
+          service_threads_manager{1, "yuzu:ServiceThreadsManager"}, system{system_} {}
 
     void SetMulticore(bool is_multi) {
         is_multicore = is_multi;
@@ -121,7 +123,7 @@ struct KernelCore::Impl {
         object_list_container.Finalize();
 
         // Ensures all service threads gracefully shutdown.
-        service_threads.clear();
+        ClearServiceThreads();
 
         next_object_id = 0;
         next_kernel_process_id = KProcess::InitialKIPIDMin;
@@ -299,12 +301,10 @@ struct KernelCore::Impl {
     // Gets the dummy KThread for the caller, allocating a new one if this is the first time
     KThread* GetHostDummyThread() {
         auto make_thread = [this]() {
-            std::lock_guard lk(dummy_thread_lock);
-            auto& thread = dummy_threads.emplace_back(std::make_unique<KThread>(system.Kernel()));
-            KAutoObject::Create(thread.get());
-            ASSERT(KThread::InitializeDummyThread(thread.get()).IsSuccess());
+            KThread* thread = KThread::Create(system.Kernel());
+            ASSERT(KThread::InitializeDummyThread(thread).IsSuccess());
             thread->SetName(fmt::format("DummyThread:{}", GetHostThreadId()));
-            return thread.get();
+            return thread;
         };
 
         thread_local KThread* saved_thread = make_thread();
@@ -629,7 +629,7 @@ struct KernelCore::Impl {
         const auto application_pool = memory_layout.GetKernelApplicationPoolRegionPhysicalExtents();
 
         // Initialize memory managers
-        memory_manager = std::make_unique<KMemoryManager>();
+        memory_manager = std::make_unique<KMemoryManager>(system);
         memory_manager->InitializeManager(KMemoryManager::Pool::Application,
                                           application_pool.GetAddress(),
                                           application_pool.GetEndAddress());
@@ -704,11 +704,31 @@ struct KernelCore::Impl {
         return port;
     }
 
+    std::weak_ptr<Kernel::ServiceThread> CreateServiceThread(KernelCore& kernel,
+                                                             const std::string& name) {
+        auto service_thread = std::make_shared<Kernel::ServiceThread>(kernel, 1, name);
+
+        service_threads_manager.QueueWork(
+            [this, service_thread]() { service_threads.emplace(service_thread); });
+
+        return service_thread;
+    }
+
+    void ReleaseServiceThread(std::weak_ptr<Kernel::ServiceThread> service_thread) {
+        if (auto strong_ptr = service_thread.lock()) {
+            service_threads_manager.QueueWork(
+                [this, strong_ptr{std::move(strong_ptr)}]() { service_threads.erase(strong_ptr); });
+        }
+    }
+
+    void ClearServiceThreads() {
+        service_threads_manager.QueueWork([this]() { service_threads.clear(); });
+    }
+
     std::mutex server_ports_lock;
     std::mutex server_sessions_lock;
     std::mutex registered_objects_lock;
     std::mutex registered_in_use_objects_lock;
-    std::mutex dummy_thread_lock;
 
     std::atomic<u32> next_object_id{0};
     std::atomic<u64> next_kernel_process_id{KProcess::InitialKIPIDMin};
@@ -759,14 +779,12 @@ struct KernelCore::Impl {
 
     // Threads used for services
     std::unordered_set<std::shared_ptr<Kernel::ServiceThread>> service_threads;
+    Common::ThreadWorker service_threads_manager;
 
     std::array<KThread*, Core::Hardware::NUM_CPU_CORES> suspend_threads;
     std::array<Core::CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES> interrupts{};
     std::array<std::unique_ptr<Kernel::KScheduler>, Core::Hardware::NUM_CPU_CORES> schedulers{};
 
-    // Specifically tracked to be automatically destroyed with kernel
-    std::vector<std::unique_ptr<KThread>> dummy_threads;
-
     bool is_multicore{};
     std::atomic_bool is_shutting_down{};
     bool is_phantom_mode_for_singlecore{};
@@ -774,6 +792,8 @@ struct KernelCore::Impl {
 
     std::array<u64, Core::Hardware::NUM_CPU_CORES> svc_ticks{};
 
+    KWorkerTaskManager worker_task_manager;
+
     // System context
     Core::System& system;
 };
@@ -1099,15 +1119,11 @@ void KernelCore::ExitSVCProfile() {
 }
 
 std::weak_ptr<Kernel::ServiceThread> KernelCore::CreateServiceThread(const std::string& name) {
-    auto service_thread = std::make_shared<Kernel::ServiceThread>(*this, 1, name);
-    impl->service_threads.emplace(service_thread);
-    return service_thread;
+    return impl->CreateServiceThread(*this, name);
 }
 
 void KernelCore::ReleaseServiceThread(std::weak_ptr<Kernel::ServiceThread> service_thread) {
-    if (auto strong_ptr = service_thread.lock()) {
-        impl->service_threads.erase(strong_ptr);
-    }
+    impl->ReleaseServiceThread(service_thread);
 }
 
 Init::KSlabResourceCounts& KernelCore::SlabResourceCounts() {
@@ -1118,6 +1134,14 @@ const Init::KSlabResourceCounts& KernelCore::SlabResourceCounts() const {
     return impl->slab_resource_counts;
 }
 
+KWorkerTaskManager& KernelCore::WorkerTaskManager() {
+    return impl->worker_task_manager;
+}
+
+const KWorkerTaskManager& KernelCore::WorkerTaskManager() const {
+    return impl->worker_task_manager;
+}
+
 bool KernelCore::IsPhantomModeForSingleCore() const {
     return impl->IsPhantomModeForSingleCore();
 }

src/core/hle/kernel/kernel.h

@@ -52,6 +52,7 @@ class KSharedMemory;
 class KSharedMemoryInfo;
 class KThread;
 class KTransferMemory;
+class KWorkerTaskManager;
 class KWritableEvent;
 class KCodeMemory;
 class PhysicalCore;
@@ -343,6 +344,12 @@ public:
     /// Gets the current slab resource counts.
     const Init::KSlabResourceCounts& SlabResourceCounts() const;
 
+    /// Gets the current worker task manager, used for dispatching KThread/KProcess tasks.
+    KWorkerTaskManager& WorkerTaskManager();
+
+    /// Gets the current worker task manager, used for dispatching KThread/KProcess tasks.
+    const KWorkerTaskManager& WorkerTaskManager() const;
+
 private:
     friend class KProcess;
     friend class KThread;

src/core/hle/kernel/physical_core.cpp

@@ -16,17 +16,25 @@ namespace Kernel {
 PhysicalCore::PhysicalCore(std::size_t core_index_, Core::System& system_, KScheduler& scheduler_,
                            Core::CPUInterrupts& interrupts_)
     : core_index{core_index_}, system{system_}, scheduler{scheduler_},
-      interrupts{interrupts_}, guard{std::make_unique<Common::SpinLock>()} {}
+      interrupts{interrupts_}, guard{std::make_unique<Common::SpinLock>()} {
+#ifdef ARCHITECTURE_x86_64
+    // TODO(bunnei): Initialization relies on a core being available. We may later replace this with
+    // a 32-bit instance of Dynarmic. This should be abstracted out to a CPU manager.
+    auto& kernel = system.Kernel();
+    arm_interface = std::make_unique<Core::ARM_Dynarmic_64>(
+        system, interrupts, kernel.IsMulticore(), kernel.GetExclusiveMonitor(), core_index);
+#else
+#error Platform not supported yet.
+#endif
+}
 
 PhysicalCore::~PhysicalCore() = default;
 
 void PhysicalCore::Initialize([[maybe_unused]] bool is_64_bit) {
 #ifdef ARCHITECTURE_x86_64
     auto& kernel = system.Kernel();
-    if (is_64_bit) {
-        arm_interface = std::make_unique<Core::ARM_Dynarmic_64>(
-            system, interrupts, kernel.IsMulticore(), kernel.GetExclusiveMonitor(), core_index);
-    } else {
+    if (!is_64_bit) {
+        // We already initialized a 64-bit core, replace with a 32-bit one.
         arm_interface = std::make_unique<Core::ARM_Dynarmic_32>(
             system, interrupts, kernel.IsMulticore(), kernel.GetExclusiveMonitor(), core_index);
     }

src/core/hle/kernel/service_thread.cpp

@@ -12,6 +12,7 @@
 #include "common/scope_exit.h"
 #include "common/thread.h"
 #include "core/hle/kernel/k_session.h"
+#include "core/hle/kernel/k_thread.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/service_thread.h"
 
@@ -50,6 +51,10 @@ ServiceThread::Impl::Impl(KernelCore& kernel, std::size_t num_threads, const std
 
             kernel.RegisterHostThread();
 
+            // Ensure the dummy thread allocated for this host thread is closed on exit.
+            auto* dummy_thread = kernel.GetCurrentEmuThread();
+            SCOPE_EXIT({ dummy_thread->Close(); });
+
             while (true) {
                 std::function<void()> task;
 

src/core/hle/kernel/svc.cpp

@@ -168,6 +168,9 @@ constexpr bool IsValidSetMemoryPermission(MemoryPermission perm) {
 
 static ResultCode SetMemoryPermission(Core::System& system, VAddr address, u64 size,
                                       MemoryPermission perm) {
+    LOG_DEBUG(Kernel_SVC, "called, address=0x{:016X}, size=0x{:X}, perm=0x{:08X", address, size,
+              perm);
+
     // Validate address / size.
     R_UNLESS(Common::IsAligned(address, PageSize), ResultInvalidAddress);
     R_UNLESS(Common::IsAligned(size, PageSize), ResultInvalidSize);
@@ -186,46 +189,33 @@ static ResultCode SetMemoryPermission(Core::System& system, VAddr address, u64 s
 }
 
 static ResultCode SetMemoryAttribute(Core::System& system, VAddr address, u64 size, u32 mask,
-                                     u32 attribute) {
+                                     u32 attr) {
     LOG_DEBUG(Kernel_SVC,
               "called, address=0x{:016X}, size=0x{:X}, mask=0x{:08X}, attribute=0x{:08X}", address,
-              size, mask, attribute);
+              size, mask, attr);
 
-    if (!Common::Is4KBAligned(address)) {
-        LOG_ERROR(Kernel_SVC, "Address not page aligned (0x{:016X})", address);
-        return ResultInvalidAddress;
-    }
+    // Validate address / size.
+    R_UNLESS(Common::IsAligned(address, PageSize), ResultInvalidAddress);
+    R_UNLESS(Common::IsAligned(size, PageSize), ResultInvalidSize);
+    R_UNLESS(size > 0, ResultInvalidSize);
+    R_UNLESS((address < address + size), ResultInvalidCurrentMemory);
 
-    if (size == 0 || !Common::Is4KBAligned(size)) {
-        LOG_ERROR(Kernel_SVC, "Invalid size (0x{:X}). Size must be non-zero and page aligned.",
-                  size);
-        return ResultInvalidAddress;
-    }
-
-    if (!IsValidAddressRange(address, size)) {
-        LOG_ERROR(Kernel_SVC, "Address range overflowed (Address: 0x{:016X}, Size: 0x{:016X})",
-                  address, size);
-        return ResultInvalidCurrentMemory;
-    }
-
-    const auto attributes{static_cast<MemoryAttribute>(mask | attribute)};
-    if (attributes != static_cast<MemoryAttribute>(mask) ||
-        (attributes | MemoryAttribute::Uncached) != MemoryAttribute::Uncached) {
-        LOG_ERROR(Kernel_SVC,
-                  "Memory attribute doesn't match the given mask (Attribute: 0x{:X}, Mask: {:X}",
-                  attribute, mask);
-        return ResultInvalidCombination;
-    }
+    // Validate the attribute and mask.
+    constexpr u32 SupportedMask = static_cast<u32>(MemoryAttribute::Uncached);
+    R_UNLESS((mask | attr) == mask, ResultInvalidCombination);
+    R_UNLESS((mask | attr | SupportedMask) == SupportedMask, ResultInvalidCombination);
 
+    // Validate that the region is in range for the current process.
     auto& page_table{system.Kernel().CurrentProcess()->PageTable()};
+    R_UNLESS(page_table.Contains(address, size), ResultInvalidCurrentMemory);
 
-    return page_table.SetMemoryAttribute(address, size, static_cast<KMemoryAttribute>(mask),
-                                         static_cast<KMemoryAttribute>(attribute));
+    // Set the memory attribute.
+    return page_table.SetMemoryAttribute(address, size, mask, attr);
 }
 
 static ResultCode SetMemoryAttribute32(Core::System& system, u32 address, u32 size, u32 mask,
-                                       u32 attribute) {
-    return SetMemoryAttribute(system, address, size, mask, attribute);
+                                       u32 attr) {
+    return SetMemoryAttribute(system, address, size, mask, attr);
 }
 
 /// Maps a memory range into a different range.
@@ -240,7 +230,7 @@ static ResultCode MapMemory(Core::System& system, VAddr dst_addr, VAddr src_addr
         return result;
     }
 
-    return page_table.Map(dst_addr, src_addr, size);
+    return page_table.MapMemory(dst_addr, src_addr, size);
 }
 
 static ResultCode MapMemory32(Core::System& system, u32 dst_addr, u32 src_addr, u32 size) {
@@ -259,7 +249,7 @@ static ResultCode UnmapMemory(Core::System& system, VAddr dst_addr, VAddr src_ad
         return result;
     }
 
-    return page_table.Unmap(dst_addr, src_addr, size);
+    return page_table.UnmapMemory(dst_addr, src_addr, size);
 }
 
 static ResultCode UnmapMemory32(Core::System& system, u32 dst_addr, u32 src_addr, u32 size) {
@@ -406,7 +396,7 @@ static ResultCode GetProcessId(Core::System& system, u64* out_process_id, Handle
     // Get the process id.
     *out_process_id = process->GetId();
 
-    return ResultInvalidHandle;
+    return ResultSuccess;
 }
 
 static ResultCode GetProcessId32(Core::System& system, u32* out_process_id_low,
@@ -655,6 +645,10 @@ static void OutputDebugString(Core::System& system, VAddr address, u64 len) {
     LOG_DEBUG(Debug_Emulated, "{}", str);
 }
 
+static void OutputDebugString32(Core::System& system, u32 address, u32 len) {
+    OutputDebugString(system, address, len);
+}
+
 /// Gets system/memory information for the current process
 static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, Handle handle,
                           u64 info_sub_id) {
@@ -1319,6 +1313,8 @@ static ResultCode SetProcessMemoryPermission(Core::System& system, Handle proces
     R_UNLESS(Common::IsAligned(size, PageSize), ResultInvalidSize);
     R_UNLESS(size > 0, ResultInvalidSize);
     R_UNLESS((address < address + size), ResultInvalidCurrentMemory);
+    R_UNLESS(address == static_cast<uintptr_t>(address), ResultInvalidCurrentMemory);
+    R_UNLESS(size == static_cast<size_t>(size), ResultInvalidCurrentMemory);
 
     // Validate the memory permission.
     R_UNLESS(IsValidProcessMemoryPermission(perm), ResultInvalidNewMemoryPermission);
@@ -1333,7 +1329,7 @@ static ResultCode SetProcessMemoryPermission(Core::System& system, Handle proces
     R_UNLESS(page_table.Contains(address, size), ResultInvalidCurrentMemory);
 
     // Set the memory permission.
-    return page_table.SetProcessMemoryPermission(address, size, ConvertToKMemoryPermission(perm));
+    return page_table.SetProcessMemoryPermission(address, size, perm);
 }
 
 static ResultCode MapProcessMemory(Core::System& system, VAddr dst_address, Handle process_handle,
@@ -1412,7 +1408,7 @@ static ResultCode UnmapProcessMemory(Core::System& system, VAddr dst_address, Ha
 }
 
 static ResultCode CreateCodeMemory(Core::System& system, Handle* out, VAddr address, size_t size) {
-    LOG_TRACE(Kernel_SVC, "called, handle_out=0x{:X}, address=0x{:X}, size=0x{:X}",
+    LOG_TRACE(Kernel_SVC, "called, handle_out={}, address=0x{:X}, size=0x{:X}",
               static_cast<void*>(out), address, size);
     // Get kernel instance.
     auto& kernel = system.Kernel();
@@ -1446,6 +1442,10 @@ static ResultCode CreateCodeMemory(Core::System& system, Handle* out, VAddr addr
     return ResultSuccess;
 }
 
+static ResultCode CreateCodeMemory32(Core::System& system, Handle* out, u32 address, u32 size) {
+    return CreateCodeMemory(system, out, address, size);
+}
+
 static ResultCode ControlCodeMemory(Core::System& system, Handle code_memory_handle, u32 operation,
                                     VAddr address, size_t size, Svc::MemoryPermission perm) {
 
@@ -1525,6 +1525,12 @@ static ResultCode ControlCodeMemory(Core::System& system, Handle code_memory_han
     return ResultSuccess;
 }
 
+static ResultCode ControlCodeMemory32(Core::System& system, Handle code_memory_handle,
+                                      u32 operation, u64 address, u64 size,
+                                      Svc::MemoryPermission perm) {
+    return ControlCodeMemory(system, code_memory_handle, operation, address, size, perm);
+}
+
 static ResultCode QueryProcessMemory(Core::System& system, VAddr memory_info_address,
                                      VAddr page_info_address, Handle process_handle,
                                      VAddr address) {
@@ -1636,7 +1642,7 @@ static ResultCode MapProcessCodeMemory(Core::System& system, Handle process_hand
         return ResultInvalidMemoryRegion;
     }
 
-    return page_table.MapProcessCodeMemory(dst_address, src_address, size);
+    return page_table.MapCodeMemory(dst_address, src_address, size);
 }
 
 static ResultCode UnmapProcessCodeMemory(Core::System& system, Handle process_handle,
@@ -1704,7 +1710,7 @@ static ResultCode UnmapProcessCodeMemory(Core::System& system, Handle process_ha
         return ResultInvalidMemoryRegion;
     }
 
-    return page_table.UnmapProcessCodeMemory(dst_address, src_address, size);
+    return page_table.UnmapCodeMemory(dst_address, src_address, size);
 }
 
 /// Exits the current process
@@ -2027,6 +2033,25 @@ static ResultCode SignalToAddress(Core::System& system, VAddr address, Svc::Sign
                                                                   count);
 }
 
+static void SynchronizePreemptionState(Core::System& system) {
+    auto& kernel = system.Kernel();
+
+    // Lock the scheduler.
+    KScopedSchedulerLock sl{kernel};
+
+    // If the current thread is pinned, unpin it.
+    KProcess* cur_process = system.Kernel().CurrentProcess();
+    const auto core_id = GetCurrentCoreId(kernel);
+
+    if (cur_process->GetPinnedThread(core_id) == GetCurrentThreadPointer(kernel)) {
+        // Clear the current thread's interrupt flag.
+        GetCurrentThread(kernel).ClearInterruptFlag();
+
+        // Unpin the current thread.
+        cur_process->UnpinCurrentThread(core_id);
+    }
+}
+
 static ResultCode SignalToAddress32(Core::System& system, u32 address, Svc::SignalType signal_type,
                                     s32 value, s32 count) {
     return SignalToAddress(system, address, signal_type, value, count);
@@ -2548,9 +2573,9 @@ struct FunctionDef {
 } // namespace
 
 static const FunctionDef SVC_Table_32[] = {
-    {0x00, nullptr, "Unknown"},
+    {0x00, nullptr, "Unknown0"},
     {0x01, SvcWrap32<SetHeapSize32>, "SetHeapSize32"},
-    {0x02, nullptr, "Unknown"},
+    {0x02, nullptr, "SetMemoryPermission32"},
     {0x03, SvcWrap32<SetMemoryAttribute32>, "SetMemoryAttribute32"},
     {0x04, SvcWrap32<MapMemory32>, "MapMemory32"},
     {0x05, SvcWrap32<UnmapMemory32>, "UnmapMemory32"},
@@ -2580,97 +2605,97 @@ static const FunctionDef SVC_Table_32[] = {
     {0x1d, SvcWrap32<SignalProcessWideKey32>, "SignalProcessWideKey32"},
     {0x1e, SvcWrap32<GetSystemTick32>, "GetSystemTick32"},
     {0x1f, SvcWrap32<ConnectToNamedPort32>, "ConnectToNamedPort32"},
-    {0x20, nullptr, "Unknown"},
+    {0x20, nullptr, "SendSyncRequestLight32"},
     {0x21, SvcWrap32<SendSyncRequest32>, "SendSyncRequest32"},
     {0x22, nullptr, "SendSyncRequestWithUserBuffer32"},
-    {0x23, nullptr, "Unknown"},
+    {0x23, nullptr, "SendAsyncRequestWithUserBuffer32"},
     {0x24, SvcWrap32<GetProcessId32>, "GetProcessId32"},
     {0x25, SvcWrap32<GetThreadId32>, "GetThreadId32"},
     {0x26, SvcWrap32<Break32>, "Break32"},
-    {0x27, nullptr, "OutputDebugString32"},
-    {0x28, nullptr, "Unknown"},
+    {0x27, SvcWrap32<OutputDebugString32>, "OutputDebugString32"},
+    {0x28, nullptr, "ReturnFromException32"},
     {0x29, SvcWrap32<GetInfo32>, "GetInfo32"},
-    {0x2a, nullptr, "Unknown"},
-    {0x2b, nullptr, "Unknown"},
+    {0x2a, nullptr, "FlushEntireDataCache32"},
+    {0x2b, nullptr, "FlushDataCache32"},
     {0x2c, SvcWrap32<MapPhysicalMemory32>, "MapPhysicalMemory32"},
     {0x2d, SvcWrap32<UnmapPhysicalMemory32>, "UnmapPhysicalMemory32"},
-    {0x2e, nullptr, "Unknown"},
-    {0x2f, nullptr, "Unknown"},
-    {0x30, nullptr, "Unknown"},
-    {0x31, nullptr, "Unknown"},
+    {0x2e, nullptr, "GetDebugFutureThreadInfo32"},
+    {0x2f, nullptr, "GetLastThreadInfo32"},
+    {0x30, nullptr, "GetResourceLimitLimitValue32"},
+    {0x31, nullptr, "GetResourceLimitCurrentValue32"},
     {0x32, SvcWrap32<SetThreadActivity32>, "SetThreadActivity32"},
     {0x33, SvcWrap32<GetThreadContext32>, "GetThreadContext32"},
     {0x34, SvcWrap32<WaitForAddress32>, "WaitForAddress32"},
     {0x35, SvcWrap32<SignalToAddress32>, "SignalToAddress32"},
-    {0x36, nullptr, "Unknown"},
-    {0x37, nullptr, "Unknown"},
-    {0x38, nullptr, "Unknown"},
-    {0x39, nullptr, "Unknown"},
-    {0x3a, nullptr, "Unknown"},
-    {0x3b, nullptr, "Unknown"},
-    {0x3c, nullptr, "Unknown"},
-    {0x3d, nullptr, "Unknown"},
-    {0x3e, nullptr, "Unknown"},
-    {0x3f, nullptr, "Unknown"},
+    {0x36, SvcWrap32<SynchronizePreemptionState>, "SynchronizePreemptionState32"},
+    {0x37, nullptr, "GetResourceLimitPeakValue32"},
+    {0x38, nullptr, "Unknown38"},
+    {0x39, nullptr, "CreateIoPool32"},
+    {0x3a, nullptr, "CreateIoRegion32"},
+    {0x3b, nullptr, "Unknown3b"},
+    {0x3c, nullptr, "KernelDebug32"},
+    {0x3d, nullptr, "ChangeKernelTraceState32"},
+    {0x3e, nullptr, "Unknown3e"},
+    {0x3f, nullptr, "Unknown3f"},
     {0x40, nullptr, "CreateSession32"},
     {0x41, nullptr, "AcceptSession32"},
-    {0x42, nullptr, "Unknown"},
+    {0x42, nullptr, "ReplyAndReceiveLight32"},
     {0x43, nullptr, "ReplyAndReceive32"},
-    {0x44, nullptr, "Unknown"},
+    {0x44, nullptr, "ReplyAndReceiveWithUserBuffer32"},
     {0x45, SvcWrap32<CreateEvent32>, "CreateEvent32"},
-    {0x46, nullptr, "Unknown"},
-    {0x47, nullptr, "Unknown"},
-    {0x48, nullptr, "Unknown"},
-    {0x49, nullptr, "Unknown"},
-    {0x4a, nullptr, "Unknown"},
-    {0x4b, nullptr, "Unknown"},
-    {0x4c, nullptr, "Unknown"},
-    {0x4d, nullptr, "Unknown"},
-    {0x4e, nullptr, "Unknown"},
-    {0x4f, nullptr, "Unknown"},
-    {0x50, nullptr, "Unknown"},
-    {0x51, nullptr, "Unknown"},
-    {0x52, nullptr, "Unknown"},
-    {0x53, nullptr, "Unknown"},
-    {0x54, nullptr, "Unknown"},
-    {0x55, nullptr, "Unknown"},
-    {0x56, nullptr, "Unknown"},
-    {0x57, nullptr, "Unknown"},
-    {0x58, nullptr, "Unknown"},
-    {0x59, nullptr, "Unknown"},
-    {0x5a, nullptr, "Unknown"},
-    {0x5b, nullptr, "Unknown"},
-    {0x5c, nullptr, "Unknown"},
-    {0x5d, nullptr, "Unknown"},
-    {0x5e, nullptr, "Unknown"},
+    {0x46, nullptr, "MapIoRegion32"},
+    {0x47, nullptr, "UnmapIoRegion32"},
+    {0x48, nullptr, "MapPhysicalMemoryUnsafe32"},
+    {0x49, nullptr, "UnmapPhysicalMemoryUnsafe32"},
+    {0x4a, nullptr, "SetUnsafeLimit32"},
+    {0x4b, SvcWrap32<CreateCodeMemory32>, "CreateCodeMemory32"},
+    {0x4c, SvcWrap32<ControlCodeMemory32>, "ControlCodeMemory32"},
+    {0x4d, nullptr, "SleepSystem32"},
+    {0x4e, nullptr, "ReadWriteRegister32"},
+    {0x4f, nullptr, "SetProcessActivity32"},
+    {0x50, nullptr, "CreateSharedMemory32"},
+    {0x51, nullptr, "MapTransferMemory32"},
+    {0x52, nullptr, "UnmapTransferMemory32"},
+    {0x53, nullptr, "CreateInterruptEvent32"},
+    {0x54, nullptr, "QueryPhysicalAddress32"},
+    {0x55, nullptr, "QueryIoMapping32"},
+    {0x56, nullptr, "CreateDeviceAddressSpace32"},
+    {0x57, nullptr, "AttachDeviceAddressSpace32"},
+    {0x58, nullptr, "DetachDeviceAddressSpace32"},
+    {0x59, nullptr, "MapDeviceAddressSpaceByForce32"},
+    {0x5a, nullptr, "MapDeviceAddressSpaceAligned32"},
+    {0x5b, nullptr, "MapDeviceAddressSpace32"},
+    {0x5c, nullptr, "UnmapDeviceAddressSpace32"},
+    {0x5d, nullptr, "InvalidateProcessDataCache32"},
+    {0x5e, nullptr, "StoreProcessDataCache32"},
     {0x5F, SvcWrap32<FlushProcessDataCache32>, "FlushProcessDataCache32"},
-    {0x60, nullptr, "Unknown"},
-    {0x61, nullptr, "Unknown"},
-    {0x62, nullptr, "Unknown"},
-    {0x63, nullptr, "Unknown"},
-    {0x64, nullptr, "Unknown"},
+    {0x60, nullptr, "StoreProcessDataCache32"},
+    {0x61, nullptr, "BreakDebugProcess32"},
+    {0x62, nullptr, "TerminateDebugProcess32"},
+    {0x63, nullptr, "GetDebugEvent32"},
+    {0x64, nullptr, "ContinueDebugEvent32"},
     {0x65, nullptr, "GetProcessList32"},
-    {0x66, nullptr, "Unknown"},
-    {0x67, nullptr, "Unknown"},
-    {0x68, nullptr, "Unknown"},
-    {0x69, nullptr, "Unknown"},
-    {0x6A, nullptr, "Unknown"},
-    {0x6B, nullptr, "Unknown"},
-    {0x6C, nullptr, "Unknown"},
-    {0x6D, nullptr, "Unknown"},
-    {0x6E, nullptr, "Unknown"},
+    {0x66, nullptr, "GetThreadList"},
+    {0x67, nullptr, "GetDebugThreadContext32"},
+    {0x68, nullptr, "SetDebugThreadContext32"},
+    {0x69, nullptr, "QueryDebugProcessMemory32"},
+    {0x6A, nullptr, "ReadDebugProcessMemory32"},
+    {0x6B, nullptr, "WriteDebugProcessMemory32"},
+    {0x6C, nullptr, "SetHardwareBreakPoint32"},
+    {0x6D, nullptr, "GetDebugThreadParam32"},
+    {0x6E, nullptr, "Unknown6E"},
     {0x6f, nullptr, "GetSystemInfo32"},
     {0x70, nullptr, "CreatePort32"},
     {0x71, nullptr, "ManageNamedPort32"},
     {0x72, nullptr, "ConnectToPort32"},
     {0x73, nullptr, "SetProcessMemoryPermission32"},
-    {0x74, nullptr, "Unknown"},
-    {0x75, nullptr, "Unknown"},
-    {0x76, nullptr, "Unknown"},
+    {0x74, nullptr, "MapProcessMemory32"},
+    {0x75, nullptr, "UnmapProcessMemory32"},
+    {0x76, nullptr, "QueryProcessMemory32"},
     {0x77, nullptr, "MapProcessCodeMemory32"},
     {0x78, nullptr, "UnmapProcessCodeMemory32"},
-    {0x79, nullptr, "Unknown"},
-    {0x7A, nullptr, "Unknown"},
+    {0x79, nullptr, "CreateProcess32"},
+    {0x7A, nullptr, "StartProcess32"},
     {0x7B, nullptr, "TerminateProcess32"},
     {0x7C, nullptr, "GetProcessInfo32"},
     {0x7D, nullptr, "CreateResourceLimit32"},
@@ -2743,7 +2768,7 @@ static const FunctionDef SVC_Table_32[] = {
 };
 
 static const FunctionDef SVC_Table_64[] = {
-    {0x00, nullptr, "Unknown"},
+    {0x00, nullptr, "Unknown0"},
     {0x01, SvcWrap64<SetHeapSize>, "SetHeapSize"},
     {0x02, SvcWrap64<SetMemoryPermission>, "SetMemoryPermission"},
     {0x03, SvcWrap64<SetMemoryAttribute>, "SetMemoryAttribute"},
@@ -2797,24 +2822,24 @@ static const FunctionDef SVC_Table_64[] = {
     {0x33, SvcWrap64<GetThreadContext>, "GetThreadContext"},
     {0x34, SvcWrap64<WaitForAddress>, "WaitForAddress"},
     {0x35, SvcWrap64<SignalToAddress>, "SignalToAddress"},
-    {0x36, nullptr, "SynchronizePreemptionState"},
-    {0x37, nullptr, "Unknown"},
-    {0x38, nullptr, "Unknown"},
-    {0x39, nullptr, "Unknown"},
-    {0x3A, nullptr, "Unknown"},
-    {0x3B, nullptr, "Unknown"},
+    {0x36, SvcWrap64<SynchronizePreemptionState>, "SynchronizePreemptionState"},
+    {0x37, nullptr, "GetResourceLimitPeakValue"},
+    {0x38, nullptr, "Unknown38"},
+    {0x39, nullptr, "CreateIoPool"},
+    {0x3A, nullptr, "CreateIoRegion"},
+    {0x3B, nullptr, "Unknown3B"},
     {0x3C, SvcWrap64<KernelDebug>, "KernelDebug"},
     {0x3D, SvcWrap64<ChangeKernelTraceState>, "ChangeKernelTraceState"},
-    {0x3E, nullptr, "Unknown"},
-    {0x3F, nullptr, "Unknown"},
+    {0x3E, nullptr, "Unknown3e"},
+    {0x3F, nullptr, "Unknown3f"},
     {0x40, nullptr, "CreateSession"},
     {0x41, nullptr, "AcceptSession"},
     {0x42, nullptr, "ReplyAndReceiveLight"},
     {0x43, nullptr, "ReplyAndReceive"},
     {0x44, nullptr, "ReplyAndReceiveWithUserBuffer"},
     {0x45, SvcWrap64<CreateEvent>, "CreateEvent"},
-    {0x46, nullptr, "Unknown"},
-    {0x47, nullptr, "Unknown"},
+    {0x46, nullptr, "MapIoRegion"},
+    {0x47, nullptr, "UnmapIoRegion"},
     {0x48, nullptr, "MapPhysicalMemoryUnsafe"},
     {0x49, nullptr, "UnmapPhysicalMemoryUnsafe"},
     {0x4A, nullptr, "SetUnsafeLimit"},
@@ -2853,7 +2878,7 @@ static const FunctionDef SVC_Table_64[] = {
     {0x6B, nullptr, "WriteDebugProcessMemory"},
     {0x6C, nullptr, "SetHardwareBreakPoint"},
     {0x6D, nullptr, "GetDebugThreadParam"},
-    {0x6E, nullptr, "Unknown"},
+    {0x6E, nullptr, "Unknown6E"},
     {0x6F, nullptr, "GetSystemInfo"},
     {0x70, nullptr, "CreatePort"},
     {0x71, nullptr, "ManageNamedPort"},

src/core/hle/kernel/svc_types.h

@@ -32,6 +32,7 @@ enum class MemoryState : u32 {
     Kernel = 0x13,
     GeneratedCode = 0x14,
     CodeOut = 0x15,
+    Coverage = 0x16,
 };
 DECLARE_ENUM_FLAG_OPERATORS(MemoryState);