common/thread: remove YieldCPU()

simply use the standard library yield()

.gitmodules

@@ -31,3 +31,6 @@
 [submodule "opus"]
     path = externals/opus
     url = https://github.com/ogniK5377/opus.git
+[submodule "soundtouch"]
+	path = externals/soundtouch
+	url = https://github.com/citra-emu/ext-soundtouch.git

.travis.yml

@@ -20,6 +20,7 @@ matrix:
       install: "./.travis/linux/deps.sh"
       script: "./.travis/linux/build.sh"
       after_success: "./.travis/linux/upload.sh"
+      cache: ccache
     - os: osx
       env: NAME="macos build"
       sudo: false
@@ -27,6 +28,20 @@ matrix:
       install: "./.travis/macos/deps.sh"
       script: "./.travis/macos/build.sh"
       after_success: "./.travis/macos/upload.sh"
+      cache: ccache
+    - os: linux
+      env: NAME="MinGW build"
+      sudo: required
+      dist: trusty
+      services: docker
+      addons:
+        apt:
+          packages:
+            - p7zip-full
+      install: "./.travis/linux-mingw/deps.sh"
+      script: "./.travis/linux-mingw/build.sh"
+      after_success: "./.travis/linux-mingw/upload.sh"
+      cache: ccache
 
 deploy:
   provider: releases
@@ -42,7 +57,3 @@ notifications:
   webhooks:
     urls:
       - https://api.yuzu-emu.org/code/travis/notify
-
-cache:
-  directories:
-    - $HOME/.ccache

.travis/common/post-upload.sh

@@ -11,6 +11,9 @@ if [ -z $TRAVIS_TAG ]; then
     RELEASE_NAME=head
 else
     RELEASE_NAME=$(echo $TRAVIS_TAG | cut -d- -f1)
+    if [ "$NAME" = "MinGW build" ]; then
+        RELEASE_NAME="${RELEASE_NAME}-mingw"
+    fi
 fi
 
 mv "$REV_NAME" $RELEASE_NAME

.travis/common/travis-ci.env

@@ -0,0 +1,12 @@
+# List of environment variables to be shared with Docker containers
+CI
+TRAVIS
+CONTINUOUS_INTEGRATION
+TRAVIS_BRANCH
+TRAVIS_BUILD_ID
+TRAVIS_BUILD_NUMBER
+TRAVIS_COMMIT
+TRAVIS_JOB_ID
+TRAVIS_JOB_NUMBER
+TRAVIS_REPO_SLUG
+TRAVIS_TAG

.travis/linux-mingw/build.sh

@@ -0,0 +1,3 @@
+#!/bin/bash -ex
+mkdir "$HOME/.ccache" || true
+docker run --env-file .travis/common/travis-ci.env -v $(pwd):/yuzu -v "$HOME/.ccache":/root/.ccache ubuntu:18.04 /bin/bash -ex /yuzu/.travis/linux-mingw/docker.sh

.travis/linux-mingw/deps.sh

@@ -0,0 +1,3 @@
+#!/bin/sh -ex
+
+docker pull ubuntu:18.04

.travis/linux-mingw/docker.sh

@@ -0,0 +1,59 @@
+#!/bin/bash -ex
+
+cd /yuzu
+MINGW_PACKAGES="sdl2-mingw-w64 qt5base-mingw-w64 qt5tools-mingw-w64 libsamplerate-mingw-w64 qt5multimedia-mingw-w64"
+apt-get update
+apt-get install -y gpg wget git python3-pip python ccache g++-mingw-w64-x86-64 gcc-mingw-w64-x86-64 mingw-w64-tools cmake
+echo 'deb http://ppa.launchpad.net/tobydox/mingw-w64/ubuntu bionic main ' > /etc/apt/sources.list.d/extras.list
+apt-key adv --keyserver keyserver.ubuntu.com --recv '72931B477E22FEFD47F8DECE02FE5F12ADDE29B2'
+apt-get update
+apt-get install -y ${MINGW_PACKAGES}
+
+# fix a problem in current MinGW headers
+wget -q https://raw.githubusercontent.com/Alexpux/mingw-w64/d0d7f784833bbb0b2d279310ddc6afb52fe47a46/mingw-w64-headers/crt/errno.h -O /usr/x86_64-w64-mingw32/include/errno.h
+# override Travis CI unreasonable ccache size
+echo 'max_size = 3.0G' > "$HOME/.ccache/ccache.conf"
+
+# Dirty hack to trick unicorn makefile into believing we are in a MINGW system
+mv /bin/uname /bin/uname1 && echo -e '#!/bin/sh\necho MINGW64' >> /bin/uname
+chmod +x /bin/uname
+
+# Dirty hack to trick unicorn makefile into believing we have cmd
+echo '' >> /bin/cmd
+chmod +x /bin/cmd
+
+mkdir build && cd build
+cmake .. -DCMAKE_TOOLCHAIN_FILE="$(pwd)/../CMakeModules/MinGWCross.cmake" -DUSE_CCACHE=ON -DYUZU_USE_BUNDLED_UNICORN=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DCMAKE_BUILD_TYPE=Release
+make -j4
+
+# Clean up the dirty hacks
+rm /bin/uname && mv /bin/uname1 /bin/uname
+rm /bin/cmd
+
+ccache -s
+
+echo "Tests skipped"
+#ctest -VV -C Release
+
+echo 'Prepare binaries...'
+cd ..
+mkdir package
+
+QT_PLATFORM_DLL_PATH='/usr/x86_64-w64-mingw32/lib/qt5/plugins/platforms/'
+find build/ -name "yuzu*.exe" -exec cp {} 'package' \;
+
+# copy Qt plugins
+mkdir package/platforms
+cp "${QT_PLATFORM_DLL_PATH}/qwindows.dll" package/platforms/
+cp -rv "${QT_PLATFORM_DLL_PATH}/../mediaservice/" package/
+cp -rv "${QT_PLATFORM_DLL_PATH}/../imageformats/" package/
+rm -f package/mediaservice/*d.dll
+
+for i in package/*.exe; do
+  # we need to process pdb here, however, cv2pdb
+  # does not work here, so we just simply strip all the debug symbols
+  x86_64-w64-mingw32-strip "${i}"
+done
+
+pip3 install pefile
+python3 .travis/linux-mingw/scan_dll.py package/*.exe "package/"

.travis/linux-mingw/scan_dll.py

@@ -0,0 +1,106 @@
+import pefile
+import sys
+import re
+import os
+import queue
+import shutil
+
+# constant definitions
+KNOWN_SYS_DLLS = ['WINMM.DLL', 'MSVCRT.DLL', 'VERSION.DLL', 'MPR.DLL',
+                  'DWMAPI.DLL', 'UXTHEME.DLL', 'DNSAPI.DLL', 'IPHLPAPI.DLL']
+# below is for Ubuntu 18.04 with specified PPA enabled, if you are using
+# other distro or different repositories, change the following accordingly
+DLL_PATH = [
+    '/usr/x86_64-w64-mingw32/bin/',
+    '/usr/x86_64-w64-mingw32/lib/',
+    '/usr/lib/gcc/x86_64-w64-mingw32/7.3-posix/'
+]
+
+missing = []
+
+
+def parse_imports(file_name):
+    results = []
+    pe = pefile.PE(file_name, fast_load=True)
+    pe.parse_data_directories()
+
+    for entry in pe.DIRECTORY_ENTRY_IMPORT:
+        current = entry.dll.decode()
+        current_u = current.upper()  # b/c Windows is often case insensitive
+        # here we filter out system dlls
+        # dll w/ names like *32.dll are likely to be system dlls
+        if current_u.upper() not in KNOWN_SYS_DLLS and not re.match(string=current_u, pattern=r'.*32\.DLL'):
+            results.append(current)
+
+    return results
+
+
+def parse_imports_recursive(file_name, path_list=[]):
+    q = queue.Queue()  # create a FIFO queue
+    # file_name can be a string or a list for the convience
+    if isinstance(file_name, str):
+        q.put(file_name)
+    elif isinstance(file_name, list):
+        for i in file_name:
+            q.put(i)
+    full_list = []
+    while q.qsize():
+        current = q.get_nowait()
+        print('> %s' % current)
+        deps = parse_imports(current)
+        # if this dll does not have any import, ignore it
+        if not deps:
+            continue
+        for dep in deps:
+            # the dependency already included in the list, skip
+            if dep in full_list:
+                continue
+            # find the requested dll in the provided paths
+            full_path = find_dll(dep)
+            if not full_path:
+                missing.append(dep)
+                continue
+            full_list.append(dep)
+            q.put(full_path)
+            path_list.append(full_path)
+    return full_list
+
+
+def find_dll(name):
+    for path in DLL_PATH:
+        for root, _, files in os.walk(path):
+            for f in files:
+                if name.lower() == f.lower():
+                    return os.path.join(root, f)
+
+
+def deploy(name, dst, dry_run=False):
+    dlls_path = []
+    parse_imports_recursive(name, dlls_path)
+    for dll_entry in dlls_path:
+        if not dry_run:
+            shutil.copy(dll_entry, dst)
+        else:
+            print('[Dry-Run] Copy %s to %s' % (dll_entry, dst))
+    print('Deploy completed.')
+    return dlls_path
+
+
+def main():
+    if len(sys.argv) < 3:
+        print('Usage: %s [files to examine ...] [target deploy directory]')
+        return 1
+    to_deploy = sys.argv[1:-1]
+    tgt_dir = sys.argv[-1]
+    if not os.path.isdir(tgt_dir):
+        print('%s is not a directory.' % tgt_dir)
+        return 1
+    print('Scanning dependencies...')
+    deploy(to_deploy, tgt_dir)
+    if missing:
+        print('Following DLLs are not found: %s' % ('\n'.join(missing)))
+    return 0
+
+
+if __name__ == '__main__':
+    main()

.travis/linux-mingw/upload.sh

@@ -0,0 +1,13 @@
+#!/bin/bash -ex
+
+. .travis/common/pre-upload.sh
+
+REV_NAME="yuzu-windows-mingw-${GITDATE}-${GITREV}"
+ARCHIVE_NAME="${REV_NAME}.tar.gz"
+COMPRESSION_FLAGS="-czvf"
+
+mkdir "$REV_NAME"
+# get around the permission issues
+cp -r package/* "$REV_NAME"
+
+. .travis/common/post-upload.sh

.travis/linux/build.sh

@@ -1,3 +1,4 @@
 #!/bin/bash -ex
 
-docker run -e CCACHE_DIR=/ccache -v $HOME/.ccache:/ccache -v $(pwd):/yuzu ubuntu:18.04 /bin/bash /yuzu/.travis/linux/docker.sh
+mkdir -p "$HOME/.ccache"
+docker run --env-file .travis/common/travis-ci.env -v $(pwd):/yuzu -v "$HOME/.ccache":/root/.ccache ubuntu:18.04 /bin/bash /yuzu/.travis/linux/docker.sh

.travis/linux/docker.sh

@@ -5,14 +5,10 @@ apt-get install --no-install-recommends -y build-essential git libqt5opengl5-dev
 
 cd /yuzu
 
-export PATH=/usr/lib/ccache:$PATH
-ln -sf /usr/bin/ccache /usr/lib/ccache/cc
-ln -sf /usr/bin/ccache /usr/lib/ccache/c++
 mkdir build && cd build
-ccache --show-stats > ccache_before
-cmake .. -DYUZU_BUILD_UNICORN=ON -DCMAKE_BUILD_TYPE=Release -G Ninja
+cmake .. -DYUZU_USE_BUNDLED_UNICORN=ON -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=/usr/lib/ccache/gcc -DCMAKE_CXX_COMPILER=/usr/lib/ccache/g++ -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -G Ninja
 ninja
-ccache --show-stats > ccache_after
-diff -U100 ccache_before ccache_after || true
+
+ccache -s
 
 ctest -VV -C Release

.travis/macos/build.sh

@@ -5,14 +5,13 @@ set -o pipefail
 export MACOSX_DEPLOYMENT_TARGET=10.12
 export Qt5_DIR=$(brew --prefix)/opt/qt5
 export UNICORNDIR=$(pwd)/externals/unicorn
+export PATH="/usr/local/opt/ccache/libexec:$PATH"
 
 mkdir build && cd build
-export PATH=/usr/local/opt/ccache/libexec:$PATH
-ccache --show-stats > ccache_before
 cmake --version
-cmake .. -DYUZU_BUILD_UNICORN=ON -DCMAKE_BUILD_TYPE=Release
+cmake .. -DYUZU_USE_BUNDLED_UNICORN=ON -DCMAKE_BUILD_TYPE=Release -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON
 make -j4
-ccache --show-stats > ccache_after
-diff -U100 ccache_before ccache_after || true
+
+ccache -s
 
 ctest -VV -C Release

CMakeLists.txt

@@ -41,6 +41,19 @@ function(check_submodules_present)
 endfunction()
 check_submodules_present()
 
+configure_file(${CMAKE_SOURCE_DIR}/dist/compatibility_list/compatibility_list.qrc
+               ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.qrc
+               COPYONLY)
+if (ENABLE_COMPATIBILITY_LIST_DOWNLOAD AND NOT EXISTS ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
+    message(STATUS "Downloading compatibility list for yuzu...")
+    file(DOWNLOAD
+        https://api.yuzu-emu.org/gamedb/
+        "${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json" SHOW_PROGRESS)
+endif()
+if (NOT EXISTS ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
+    file(WRITE ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json "")
+endif()
+
 # Detect current compilation architecture and create standard definitions
 # =======================================================================
 
@@ -256,10 +269,18 @@ if (YUZU_USE_BUNDLED_UNICORN)
 
         find_package(PythonInterp 2.7 REQUIRED)
 
-        add_custom_command(OUTPUT ${LIBUNICORN_LIBRARY}
-            COMMAND ${CMAKE_COMMAND} -E env UNICORN_ARCHS="aarch64" PYTHON="${PYTHON_EXECUTABLE}" /bin/sh make.sh macos-universal-no
-            WORKING_DIRECTORY ${UNICORN_PREFIX}
-        )
+        if (MINGW)
+            add_custom_command(OUTPUT ${LIBUNICORN_LIBRARY}
+                COMMAND ${CMAKE_COMMAND} -E env UNICORN_ARCHS="aarch64" PYTHON="${PYTHON_EXECUTABLE}" /bin/sh make.sh cross-win64
+                WORKING_DIRECTORY ${UNICORN_PREFIX}
+            )
+        else()
+            add_custom_command(OUTPUT ${LIBUNICORN_LIBRARY}
+                COMMAND ${CMAKE_COMMAND} -E env UNICORN_ARCHS="aarch64" PYTHON="${PYTHON_EXECUTABLE}" /bin/sh make.sh macos-universal-no
+                WORKING_DIRECTORY ${UNICORN_PREFIX}
+            )
+        endif()
+
         # ALL makes this custom target build every time
         # but it won't actually build if LIBUNICORN_LIBRARY is up to date
         add_custom_target(unicorn-build ALL
@@ -273,6 +294,7 @@ endif()
 
 if (UNICORN_FOUND)
     add_library(unicorn INTERFACE)
+    add_dependencies(unicorn unicorn-build)
     target_link_libraries(unicorn INTERFACE "${LIBUNICORN_LIBRARY}")
     target_include_directories(unicorn INTERFACE "${LIBUNICORN_INCLUDE_DIR}")
 else()
@@ -418,6 +440,9 @@ enable_testing()
 add_subdirectory(externals)
 add_subdirectory(src)
 
+# Set yuzu project as default StartUp Project in Visual Studio
+set_property(DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR} PROPERTY VS_STARTUP_PROJECT yuzu)
+
 
 # Installation instructions
 # =========================

CMakeModules/MinGWCross.cmake

@@ -0,0 +1,54 @@
+set(MINGW_PREFIX   /usr/x86_64-w64-mingw32/)
+set(CMAKE_SYSTEM_NAME               Windows)
+set(CMAKE_SYSTEM_PROCESSOR           x86_64)
+# Actually a hack, w/o this will cause some strange errors
+set(CMAKE_HOST_WIN32                 TRUE)
+
+
+set(CMAKE_FIND_ROOT_PATH            ${MINGW_PREFIX})
+set(SDL2_PATH                       ${MINGW_PREFIX})
+set(MINGW_TOOL_PREFIX               ${CMAKE_SYSTEM_PROCESSOR}-w64-mingw32-)
+
+# Specify the cross compiler
+set(CMAKE_C_COMPILER            ${MINGW_TOOL_PREFIX}gcc-posix)
+set(CMAKE_CXX_COMPILER          ${MINGW_TOOL_PREFIX}g++-posix)
+set(CMAKE_RC_COMPILER           ${MINGW_TOOL_PREFIX}windres)
+
+# Mingw tools
+set(STRIP                       ${MINGW_TOOL_PREFIX}strip)
+set(WINDRES                     ${MINGW_TOOL_PREFIX}windres)
+set(ENV{PKG_CONFIG}             ${MINGW_TOOL_PREFIX}pkg-config)
+
+# ccache wrapper
+option(USE_CCACHE "Use ccache for compilation" OFF)
+if(USE_CCACHE)
+    find_program(CCACHE ccache)
+    if(CCACHE)
+        message(STATUS "Using ccache found in PATH")
+        set_property(GLOBAL PROPERTY RULE_LAUNCH_COMPILE ${CCACHE})
+        set_property(GLOBAL PROPERTY RULE_LAUNCH_LINK ${CCACHE})
+    else(CCACHE)
+        message(WARNING "USE_CCACHE enabled, but no ccache found")
+    endif(CCACHE)
+endif(USE_CCACHE)
+
+# Search for programs in the build host directories
+set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER)
+
+
+# Echo modified cmake vars to screen for debugging purposes
+if(NOT DEFINED ENV{MINGW_DEBUG_INFO})
+        message("")
+        message("Custom cmake vars: (blank = system default)")
+        message("-----------------------------------------")
+        message("* CMAKE_C_COMPILER                     : ${CMAKE_C_COMPILER}")
+        message("* CMAKE_CXX_COMPILER                   : ${CMAKE_CXX_COMPILER}")
+        message("* CMAKE_RC_COMPILER                    : ${CMAKE_RC_COMPILER}")
+        message("* WINDRES                              : ${WINDRES}")
+        message("* ENV{PKG_CONFIG}                      : $ENV{PKG_CONFIG}")
+        message("* STRIP                                : ${STRIP}")
+        message("* USE_CCACHE                           : ${USE_CCACHE}")
+        message("")
+        # So that the debug info only appears once
+        set(ENV{MINGW_DEBUG_INFO} SHOWN)
+endif()

appveyor.yml

@@ -41,9 +41,9 @@ before_build:
   - ps: |
         if ($env:BUILD_TYPE -eq 'msvc') {
           # redirect stderr and change the exit code to prevent powershell from cancelling the build if cmake prints a warning
-          cmd /C 'cmake -G "Visual Studio 15 2017 Win64" -DYUZU_USE_BUNDLED_QT=1 -DYUZU_USE_BUNDLED_SDL2=1 -DYUZU_USE_BUNDLED_UNICORN=1 .. 2>&1 && exit 0'
+          cmd /C 'cmake -G "Visual Studio 15 2017 Win64" -DYUZU_USE_BUNDLED_QT=1 -DYUZU_USE_BUNDLED_SDL2=1 -DYUZU_USE_BUNDLED_UNICORN=1 -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON .. 2>&1 && exit 0'
         } else {
-          C:\msys64\usr\bin\bash.exe -lc "cmake -G 'MSYS Makefiles' -DYUZU_BUILD_UNICORN=1 -DCMAKE_BUILD_TYPE=Release .. 2>&1"
+          C:\msys64\usr\bin\bash.exe -lc "cmake -G 'MSYS Makefiles' -DYUZU_BUILD_UNICORN=1 -DCMAKE_BUILD_TYPE=Release -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON .. 2>&1"
         }
   - cd ..
 
@@ -162,10 +162,6 @@ artifacts:
   - path: $(BUILD_ZIP)
     name: build
     type: zip
-  - path: $(BUILD_SYMBOLS)
-    name: debugsymbols
-  - path: $(BUILD_UPDATE)
-    name: update
 
 deploy:
   provider: GitHub

dist/compatibility_list/compatibility_list.qrc

@@ -0,0 +1,5 @@
+<RCC>
+  <qresource prefix="compatibility_list">
+      <file>compatibility_list.json</file>
+  </qresource>
+</RCC>

externals/CMakeLists.txt

@@ -43,10 +43,16 @@ target_include_directories(mbedtls PUBLIC ./mbedtls/include)
 add_library(microprofile INTERFACE)
 target_include_directories(microprofile INTERFACE ./microprofile)
 
+# Open Source Archives
+add_subdirectory(open_source_archives EXCLUDE_FROM_ALL)
+
 # Unicorn
 add_library(unicorn-headers INTERFACE)
 target_include_directories(unicorn-headers INTERFACE ./unicorn/include)
 
+# SoundTouch
+add_subdirectory(soundtouch)
+
 # Xbyak
 if (ARCHITECTURE_x86_64)
     # Defined before "dynarmic" above

externals/open_source_archives/CMakeLists.txt

@@ -0,0 +1,16 @@
+add_library(open_source_archives
+    src/FontChineseSimplified.cpp
+    src/FontChineseTraditional.cpp
+    src/FontExtendedChineseSimplified.cpp
+    src/FontKorean.cpp
+    src/FontNintendoExtended.cpp
+    src/FontStandard.cpp
+    include/FontChineseSimplified.h
+    include/FontChineseTraditional.h
+    include/FontExtendedChineseSimplified.h
+    include/FontKorean.h
+    include/FontNintendoExtended.h
+    include/FontStandard.h
+)
+
+target_include_directories(open_source_archives PUBLIC include)

externals/open_source_archives/Readme.md

@@ -0,0 +1,4 @@
+These files were generated by https://github.com/FearlessTobi/yuzu_system_archives at git commit 0a24b0c9f38d71fb2c4bba5645a39029e539a5ec. To generate the files use the run.sh inside that repository.  
+
+The follwing system archives are currently included:
+ - JPN/EUR/USA System Font

externals/open_source_archives/include/FontChineseSimplified.h

@@ -0,0 +1,6 @@
+#pragma once
+
+#include <array>
+
+extern const std::array<unsigned char, 217276> FontChineseSimplified;
+

externals/open_source_archives/include/FontChineseTraditional.h

@@ -0,0 +1,6 @@
+#pragma once
+
+#include <array>
+
+extern const std::array<unsigned char, 222236> FontChineseTraditional;
+

externals/open_source_archives/include/FontExtendedChineseSimplified.h

@@ -0,0 +1,6 @@
+#pragma once
+
+#include <array>
+
+extern const std::array<unsigned char, 293516> FontExtendedChineseSimplified;
+

externals/open_source_archives/include/FontKorean.h

@@ -0,0 +1,6 @@
+#pragma once
+
+#include <array>
+
+extern const std::array<unsigned char, 217276> FontKorean;
+

externals/open_source_archives/include/FontNintendoExtended.h

@@ -0,0 +1,6 @@
+#pragma once
+
+#include <array>
+
+extern const std::array<unsigned char, 172064> FontNintendoExtended;
+

externals/open_source_archives/include/FontStandard.h

@@ -0,0 +1,6 @@
+#pragma once
+
+#include <array>
+
+extern const std::array<unsigned char, 217276> FontStandard;
+

src/audio_core/CMakeLists.txt

@@ -17,6 +17,8 @@ add_library(audio_core STATIC
     sink_stream.h
     stream.cpp
     stream.h
+    time_stretch.cpp
+    time_stretch.h
 
     $<$<BOOL:${ENABLE_CUBEB}>:cubeb_sink.cpp cubeb_sink.h>
 )
@@ -24,6 +26,7 @@ add_library(audio_core STATIC
 create_target_directory_groups(audio_core)
 
 target_link_libraries(audio_core PUBLIC common core)
+target_link_libraries(audio_core PRIVATE SoundTouch)
 
 if(ENABLE_CUBEB)
     target_link_libraries(audio_core PRIVATE cubeb)

src/audio_core/algorithm/filter.cpp

@@ -35,12 +35,12 @@ Filter::Filter(double a0, double a1, double a2, double b0, double b1, double b2)
     : a1(a1 / a0), a2(a2 / a0), b0(b0 / a0), b1(b1 / a0), b2(b2 / a0) {}
 
 void Filter::Process(std::vector<s16>& signal) {
-    const size_t num_frames = signal.size() / 2;
-    for (size_t i = 0; i < num_frames; i++) {
+    const std::size_t num_frames = signal.size() / 2;
+    for (std::size_t i = 0; i < num_frames; i++) {
         std::rotate(in.begin(), in.end() - 1, in.end());
         std::rotate(out.begin(), out.end() - 1, out.end());
 
-        for (size_t ch = 0; ch < channel_count; ch++) {
+        for (std::size_t ch = 0; ch < channel_count; ch++) {
             in[0][ch] = signal[i * channel_count + ch];
 
             out[0][ch] = b0 * in[0][ch] + b1 * in[1][ch] + b2 * in[2][ch] - a1 * out[1][ch] -
@@ -54,14 +54,14 @@ void Filter::Process(std::vector<s16>& signal) {
 /// Calculates the appropriate Q for each biquad in a cascading filter.
 /// @param total_count The total number of biquads to be cascaded.
 /// @param index 0-index of the biquad to calculate the Q value for.
-static double CascadingBiquadQ(size_t total_count, size_t index) {
+static double CascadingBiquadQ(std::size_t total_count, std::size_t index) {
     const double pole = M_PI * (2 * index + 1) / (4.0 * total_count);
     return 1.0 / (2.0 * std::cos(pole));
 }
 
-CascadingFilter CascadingFilter::LowPass(double cutoff, size_t cascade_size) {
+CascadingFilter CascadingFilter::LowPass(double cutoff, std::size_t cascade_size) {
     std::vector<Filter> cascade(cascade_size);
-    for (size_t i = 0; i < cascade_size; i++) {
+    for (std::size_t i = 0; i < cascade_size; i++) {
         cascade[i] = Filter::LowPass(cutoff, CascadingBiquadQ(cascade_size, i));
     }
     return CascadingFilter{std::move(cascade)};

src/audio_core/algorithm/filter.h

@@ -30,7 +30,7 @@ public:
     void Process(std::vector<s16>& signal);
 
 private:
-    static constexpr size_t channel_count = 2;
+    static constexpr std::size_t channel_count = 2;
 
     /// Coefficients are in normalized form (a0 = 1.0).
     double a1, a2, b0, b1, b2;
@@ -46,7 +46,7 @@ public:
     /// Creates a cascading low-pass filter.
     /// @param cutoff Determines the cutoff frequency. A value from 0.0 to 1.0.
     /// @param cascade_size Number of biquads in cascade.
-    static CascadingFilter LowPass(double cutoff, size_t cascade_size);
+    static CascadingFilter LowPass(double cutoff, std::size_t cascade_size);
 
     /// Passthrough.
     CascadingFilter();

src/audio_core/algorithm/interpolate.cpp

@@ -14,7 +14,7 @@
 namespace AudioCore {
 
 /// The Lanczos kernel
-static double Lanczos(size_t a, double x) {
+static double Lanczos(std::size_t a, double x) {
     if (x == 0.0)
         return 1.0;
     const double px = M_PI * x;
@@ -37,15 +37,15 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input,
     }
     state.nyquist.Process(input);
 
-    constexpr size_t taps = InterpolationState::lanczos_taps;
-    const size_t num_frames = input.size() / 2;
+    constexpr std::size_t taps = InterpolationState::lanczos_taps;
+    const std::size_t num_frames = input.size() / 2;
 
     std::vector<s16> output;
-    output.reserve(static_cast<size_t>(input.size() / ratio + 4));
+    output.reserve(static_cast<std::size_t>(input.size() / ratio + 4));
 
     double& pos = state.position;
     auto& h = state.history;
-    for (size_t i = 0; i < num_frames; ++i) {
+    for (std::size_t i = 0; i < num_frames; ++i) {
         std::rotate(h.begin(), h.end() - 1, h.end());
         h[0][0] = input[i * 2 + 0];
         h[0][1] = input[i * 2 + 1];
@@ -53,7 +53,7 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input,
         while (pos <= 1.0) {
             double l = 0.0;
             double r = 0.0;
-            for (size_t j = 0; j < h.size(); j++) {
+            for (std::size_t j = 0; j < h.size(); j++) {
                 l += Lanczos(taps, pos + j - taps + 1) * h[j][0];
                 r += Lanczos(taps, pos + j - taps + 1) * h[j][1];
             }

src/audio_core/algorithm/interpolate.h

@@ -12,8 +12,8 @@
 namespace AudioCore {
 
 struct InterpolationState {
-    static constexpr size_t lanczos_taps = 4;
-    static constexpr size_t history_size = lanczos_taps * 2 - 1;
+    static constexpr std::size_t lanczos_taps = 4;
+    static constexpr std::size_t history_size = lanczos_taps * 2 - 1;
 
     double current_ratio = 0.0;
     CascadingFilter nyquist;

src/audio_core/audio_out.cpp

@@ -39,7 +39,8 @@ StreamPtr AudioOut::OpenStream(u32 sample_rate, u32 num_channels, std::string&&
         sink->AcquireSinkStream(sample_rate, num_channels, name), std::move(name));
 }
 
-std::vector<Buffer::Tag> AudioOut::GetTagsAndReleaseBuffers(StreamPtr stream, size_t max_count) {
+std::vector<Buffer::Tag> AudioOut::GetTagsAndReleaseBuffers(StreamPtr stream,
+                                                            std::size_t max_count) {
     return stream->GetTagsAndReleaseBuffers(max_count);
 }
 

src/audio_core/audio_out.h

@@ -25,7 +25,7 @@ public:
                          Stream::ReleaseCallback&& release_callback);
 
     /// Returns a vector of recently released buffers specified by tag for the specified stream
-    std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(StreamPtr stream, size_t max_count);
+    std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(StreamPtr stream, std::size_t max_count);
 
     /// Starts an audio stream for playback
     void StartStream(StreamPtr stream);

src/audio_core/audio_renderer.cpp

@@ -3,9 +3,12 @@
 // Refer to the license.txt file included.
 
 #include "audio_core/algorithm/interpolate.h"
+#include "audio_core/audio_out.h"
 #include "audio_core/audio_renderer.h"
+#include "audio_core/codec.h"
 #include "common/assert.h"
 #include "common/logging/log.h"
+#include "core/hle/kernel/event.h"
 #include "core/memory.h"
 
 namespace AudioCore {
@@ -13,20 +16,57 @@ namespace AudioCore {
 constexpr u32 STREAM_SAMPLE_RATE{48000};
 constexpr u32 STREAM_NUM_CHANNELS{2};
 
+class AudioRenderer::VoiceState {
+public:
+    bool IsPlaying() const {
+        return is_in_use && info.play_state == PlayState::Started;
+    }
+
+    const VoiceOutStatus& GetOutStatus() const {
+        return out_status;
+    }
+
+    const VoiceInfo& GetInfo() const {
+        return info;
+    }
+
+    VoiceInfo& Info() {
+        return info;
+    }
+
+    void SetWaveIndex(std::size_t index);
+    std::vector<s16> DequeueSamples(std::size_t sample_count);
+    void UpdateState();
+    void RefreshBuffer();
+
+private:
+    bool is_in_use{};
+    bool is_refresh_pending{};
+    std::size_t wave_index{};
+    std::size_t offset{};
+    Codec::ADPCMState adpcm_state{};
+    InterpolationState interp_state{};
+    std::vector<s16> samples;
+    VoiceOutStatus out_status{};
+    VoiceInfo info{};
+};
+
 AudioRenderer::AudioRenderer(AudioRendererParameter params,
                              Kernel::SharedPtr<Kernel::Event> buffer_event)
     : worker_params{params}, buffer_event{buffer_event}, voices(params.voice_count) {
 
-    audio_core = std::make_unique<AudioCore::AudioOut>();
-    stream = audio_core->OpenStream(STREAM_SAMPLE_RATE, STREAM_NUM_CHANNELS, "AudioRenderer",
-                                    [=]() { buffer_event->Signal(); });
-    audio_core->StartStream(stream);
+    audio_out = std::make_unique<AudioCore::AudioOut>();
+    stream = audio_out->OpenStream(STREAM_SAMPLE_RATE, STREAM_NUM_CHANNELS, "AudioRenderer",
+                                   [=]() { buffer_event->Signal(); });
+    audio_out->StartStream(stream);
 
     QueueMixedBuffer(0);
     QueueMixedBuffer(1);
     QueueMixedBuffer(2);
 }
 
+AudioRenderer::~AudioRenderer() = default;
+
 u32 AudioRenderer::GetSampleRate() const {
     return worker_params.sample_rate;
 }
@@ -52,8 +92,8 @@ std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_
                 memory_pool_count * sizeof(MemoryPoolInfo));
 
     // Copy VoiceInfo structs
-    size_t offset{sizeof(UpdateDataHeader) + config.behavior_size + config.memory_pools_size +
-                  config.voice_resource_size};
+    std::size_t offset{sizeof(UpdateDataHeader) + config.behavior_size + config.memory_pools_size +
+                       config.voice_resource_size};
     for (auto& voice : voices) {
         std::memcpy(&voice.Info(), input_params.data() + offset, sizeof(VoiceInfo));
         offset += sizeof(VoiceInfo);
@@ -72,7 +112,7 @@ std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_
 
     // Update memory pool state
     std::vector<MemoryPoolEntry> memory_pool(memory_pool_count);
-    for (size_t index = 0; index < memory_pool.size(); ++index) {
+    for (std::size_t index = 0; index < memory_pool.size(); ++index) {
         if (mem_pool_info[index].pool_state == MemoryPoolStates::RequestAttach) {
             memory_pool[index].state = MemoryPoolStates::Attached;
         } else if (mem_pool_info[index].pool_state == MemoryPoolStates::RequestDetach) {
@@ -93,7 +133,7 @@ std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_
                 response_data.memory_pools_size);
 
     // Copy output voice status
-    size_t voice_out_status_offset{sizeof(UpdateDataHeader) + response_data.memory_pools_size};
+    std::size_t voice_out_status_offset{sizeof(UpdateDataHeader) + response_data.memory_pools_size};
     for (const auto& voice : voices) {
         std::memcpy(output_params.data() + voice_out_status_offset, &voice.GetOutStatus(),
                     sizeof(VoiceOutStatus));
@@ -103,12 +143,12 @@ std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_
     return output_params;
 }
 
-void AudioRenderer::VoiceState::SetWaveIndex(size_t index) {
+void AudioRenderer::VoiceState::SetWaveIndex(std::size_t index) {
     wave_index = index & 3;
     is_refresh_pending = true;
 }
 
-std::vector<s16> AudioRenderer::VoiceState::DequeueSamples(size_t sample_count) {
+std::vector<s16> AudioRenderer::VoiceState::DequeueSamples(std::size_t sample_count) {
     if (!IsPlaying()) {
         return {};
     }
@@ -117,9 +157,9 @@ std::vector<s16> AudioRenderer::VoiceState::DequeueSamples(size_t sample_count)
         RefreshBuffer();
     }
 
-    const size_t max_size{samples.size() - offset};
-    const size_t dequeue_offset{offset};
-    size_t size{sample_count * STREAM_NUM_CHANNELS};
+    const std::size_t max_size{samples.size() - offset};
+    const std::size_t dequeue_offset{offset};
+    std::size_t size{sample_count * STREAM_NUM_CHANNELS};
     if (size > max_size) {
         size = max_size;
     }
@@ -184,7 +224,7 @@ void AudioRenderer::VoiceState::RefreshBuffer() {
     case 1:
         // 1 channel is upsampled to 2 channel
         samples.resize(new_samples.size() * 2);
-        for (size_t index = 0; index < new_samples.size(); ++index) {
+        for (std::size_t index = 0; index < new_samples.size(); ++index) {
             samples[index * 2] = new_samples[index];
             samples[index * 2 + 1] = new_samples[index];
         }
@@ -210,7 +250,7 @@ static constexpr s16 ClampToS16(s32 value) {
 }
 
 void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) {
-    constexpr size_t BUFFER_SIZE{512};
+    constexpr std::size_t BUFFER_SIZE{512};
     std::vector<s16> buffer(BUFFER_SIZE * stream->GetNumChannels());
 
     for (auto& voice : voices) {
@@ -218,7 +258,7 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) {
             continue;
         }
 
-        size_t offset{};
+        std::size_t offset{};
         s64 samples_remaining{BUFFER_SIZE};
         while (samples_remaining > 0) {
             const std::vector<s16> samples{voice.DequeueSamples(samples_remaining)};
@@ -236,11 +276,11 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) {
             }
         }
     }
-    audio_core->QueueBuffer(stream, tag, std::move(buffer));
+    audio_out->QueueBuffer(stream, tag, std::move(buffer));
 }
 
 void AudioRenderer::ReleaseAndQueueBuffers() {
-    const auto released_buffers{audio_core->GetTagsAndReleaseBuffers(stream, 2)};
+    const auto released_buffers{audio_out->GetTagsAndReleaseBuffers(stream, 2)};
     for (const auto& tag : released_buffers) {
         QueueMixedBuffer(tag);
     }

src/audio_core/audio_renderer.h

@@ -8,16 +8,20 @@
 #include <memory>
 #include <vector>
 
-#include "audio_core/algorithm/interpolate.h"
-#include "audio_core/audio_out.h"
-#include "audio_core/codec.h"
 #include "audio_core/stream.h"
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/swap.h"
-#include "core/hle/kernel/event.h"
+#include "core/hle/kernel/object.h"
+
+namespace Kernel {
+class Event;
+}
 
 namespace AudioCore {
 
+class AudioOut;
+
 enum class PlayState : u8 {
     Started = 0,
     Stopped = 1,
@@ -158,6 +162,8 @@ static_assert(sizeof(UpdateDataHeader) == 0x40, "UpdateDataHeader has wrong size
 class AudioRenderer {
 public:
     AudioRenderer(AudioRendererParameter params, Kernel::SharedPtr<Kernel::Event> buffer_event);
+    ~AudioRenderer();
+
     std::vector<u8> UpdateAudioRenderer(const std::vector<u8>& input_params);
     void QueueMixedBuffer(Buffer::Tag tag);
     void ReleaseAndQueueBuffers();
@@ -166,45 +172,12 @@ public:
     u32 GetMixBufferCount() const;
 
 private:
-    class VoiceState {
-    public:
-        bool IsPlaying() const {
-            return is_in_use && info.play_state == PlayState::Started;
-        }
-
-        const VoiceOutStatus& GetOutStatus() const {
-            return out_status;
-        }
-
-        const VoiceInfo& GetInfo() const {
-            return info;
-        }
-
-        VoiceInfo& Info() {
-            return info;
-        }
-
-        void SetWaveIndex(size_t index);
-        std::vector<s16> DequeueSamples(size_t sample_count);
-        void UpdateState();
-        void RefreshBuffer();
-
-    private:
-        bool is_in_use{};
-        bool is_refresh_pending{};
-        size_t wave_index{};
-        size_t offset{};
-        Codec::ADPCMState adpcm_state{};
-        InterpolationState interp_state{};
-        std::vector<s16> samples;
-        VoiceOutStatus out_status{};
-        VoiceInfo info{};
-    };
+    class VoiceState;
 
     AudioRendererParameter worker_params;
     Kernel::SharedPtr<Kernel::Event> buffer_event;
     std::vector<VoiceState> voices;
-    std::unique_ptr<AudioCore::AudioOut> audio_core;
+    std::unique_ptr<AudioOut> audio_out;
     AudioCore::StreamPtr stream;
 };
 

src/audio_core/codec.cpp

@@ -8,27 +8,27 @@
 
 namespace AudioCore::Codec {
 
-std::vector<s16> DecodeADPCM(const u8* const data, size_t size, const ADPCM_Coeff& coeff,
+std::vector<s16> DecodeADPCM(const u8* const data, std::size_t size, const ADPCM_Coeff& coeff,
                              ADPCMState& state) {
     // GC-ADPCM with scale factor and variable coefficients.
     // Frames are 8 bytes long containing 14 samples each.
     // Samples are 4 bits (one nibble) long.
 
-    constexpr size_t FRAME_LEN = 8;
-    constexpr size_t SAMPLES_PER_FRAME = 14;
+    constexpr std::size_t FRAME_LEN = 8;
+    constexpr std::size_t SAMPLES_PER_FRAME = 14;
     constexpr std::array<int, 16> SIGNED_NIBBLES = {
         {0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1}};
 
-    const size_t sample_count = (size / FRAME_LEN) * SAMPLES_PER_FRAME;
-    const size_t ret_size =
+    const std::size_t sample_count = (size / FRAME_LEN) * SAMPLES_PER_FRAME;
+    const std::size_t ret_size =
         sample_count % 2 == 0 ? sample_count : sample_count + 1; // Ensure multiple of two.
     std::vector<s16> ret(ret_size);
 
     int yn1 = state.yn1, yn2 = state.yn2;
 
-    const size_t NUM_FRAMES =
+    const std::size_t NUM_FRAMES =
         (sample_count + (SAMPLES_PER_FRAME - 1)) / SAMPLES_PER_FRAME; // Round up.
-    for (size_t framei = 0; framei < NUM_FRAMES; framei++) {
+    for (std::size_t framei = 0; framei < NUM_FRAMES; framei++) {
         const int frame_header = data[framei * FRAME_LEN];
         const int scale = 1 << (frame_header & 0xF);
         const int idx = (frame_header >> 4) & 0x7;
@@ -53,9 +53,9 @@ std::vector<s16> DecodeADPCM(const u8* const data, size_t size, const ADPCM_Coef
             return static_cast<s16>(val);
         };
 
-        size_t outputi = framei * SAMPLES_PER_FRAME;
-        size_t datai = framei * FRAME_LEN + 1;
-        for (size_t i = 0; i < SAMPLES_PER_FRAME && outputi < sample_count; i += 2) {
+        std::size_t outputi = framei * SAMPLES_PER_FRAME;
+        std::size_t datai = framei * FRAME_LEN + 1;
+        for (std::size_t i = 0; i < SAMPLES_PER_FRAME && outputi < sample_count; i += 2) {
             const s16 sample1 = decode_sample(SIGNED_NIBBLES[data[datai] >> 4]);
             ret[outputi] = sample1;
             outputi++;

src/audio_core/codec.h

@@ -38,7 +38,7 @@ using ADPCM_Coeff = std::array<s16, 16>;
  * @param state ADPCM state, this is updated with new state
  * @return Decoded stereo signed PCM16 data, sample_count in length
  */
-std::vector<s16> DecodeADPCM(const u8* const data, size_t size, const ADPCM_Coeff& coeff,
+std::vector<s16> DecodeADPCM(const u8* const data, std::size_t size, const ADPCM_Coeff& coeff,
                              ADPCMState& state);
 
 }; // namespace AudioCore::Codec

src/audio_core/cubeb_sink.cpp

@@ -3,27 +3,23 @@
 // Refer to the license.txt file included.
 
 #include <algorithm>
+#include <atomic>
 #include <cstring>
-#include <mutex>
-
 #include "audio_core/cubeb_sink.h"
 #include "audio_core/stream.h"
+#include "audio_core/time_stretch.h"
 #include "common/logging/log.h"
+#include "common/ring_buffer.h"
+#include "core/settings.h"
 
 namespace AudioCore {
 
-class SinkStreamImpl final : public SinkStream {
+class CubebSinkStream final : public SinkStream {
 public:
-    SinkStreamImpl(cubeb* ctx, u32 sample_rate, u32 num_channels_, cubeb_devid output_device,
-                   const std::string& name)
-        : ctx{ctx}, num_channels{num_channels_} {
-
-        if (num_channels == 6) {
-            // 6-channel audio does not seem to work with cubeb + SDL, so we downsample this to 2
-            // channel for now
-            is_6_channel = true;
-            num_channels = 2;
-        }
+    CubebSinkStream(cubeb* ctx, u32 sample_rate, u32 num_channels_, cubeb_devid output_device,
+                    const std::string& name)
+        : ctx{ctx}, num_channels{std::min(num_channels_, 2u)}, time_stretch{sample_rate,
+                                                                            num_channels} {
 
         cubeb_stream_params params{};
         params.rate = sample_rate;
@@ -38,7 +34,7 @@ public:
 
         if (cubeb_stream_init(ctx, &stream_backend, name.c_str(), nullptr, nullptr, output_device,
                               &params, std::max(512u, minimum_latency),
-                              &SinkStreamImpl::DataCallback, &SinkStreamImpl::StateCallback,
+                              &CubebSinkStream::DataCallback, &CubebSinkStream::StateCallback,
                               this) != CUBEB_OK) {
             LOG_CRITICAL(Audio_Sink, "Error initializing cubeb stream");
             return;
@@ -50,7 +46,7 @@ public:
         }
     }
 
-    ~SinkStreamImpl() {
+    ~CubebSinkStream() {
         if (!ctx) {
             return;
         }
@@ -62,27 +58,32 @@ public:
         cubeb_stream_destroy(stream_backend);
     }
 
-    void EnqueueSamples(u32 num_channels, const std::vector<s16>& samples) override {
-        if (!ctx) {
+    void EnqueueSamples(u32 source_num_channels, const std::vector<s16>& samples) override {
+        if (source_num_channels > num_channels) {
+            // Downsample 6 channels to 2
+            std::vector<s16> buf;
+            buf.reserve(samples.size() * num_channels / source_num_channels);
+            for (std::size_t i = 0; i < samples.size(); i += source_num_channels) {
+                for (std::size_t ch = 0; ch < num_channels; ch++) {
+                    buf.push_back(samples[i + ch]);
+                }
+            }
+            queue.Push(buf);
             return;
         }
 
-        std::lock_guard lock{queue_mutex};
+        queue.Push(samples);
+    }
 
-        queue.reserve(queue.size() + samples.size() * GetNumChannels());
+    std::size_t SamplesInQueue(u32 num_channels) const override {
+        if (!ctx)
+            return 0;
 
-        if (is_6_channel) {
-            // Downsample 6 channels to 2
-            const size_t sample_count_copy_size = samples.size() * 2;
-            queue.reserve(sample_count_copy_size);
-            for (size_t i = 0; i < samples.size(); i += num_channels) {
-                queue.push_back(samples[i]);
-                queue.push_back(samples[i + 1]);
-            }
-        } else {
-            // Copy as-is
-            std::copy(samples.begin(), samples.end(), std::back_inserter(queue));
-        }
+        return queue.Size() / num_channels;
+    }
+
+    void Flush() override {
+        should_flush = true;
     }
 
     u32 GetNumChannels() const {
@@ -95,10 +96,11 @@ private:
     cubeb* ctx{};
     cubeb_stream* stream_backend{};
     u32 num_channels{};
-    bool is_6_channel{};
 
-    std::mutex queue_mutex;
-    std::vector<s16> queue;
+    Common::RingBuffer<s16, 0x10000> queue;
+    std::array<s16, 2> last_frame;
+    std::atomic<bool> should_flush{};
+    TimeStretcher time_stretch;
 
     static long DataCallback(cubeb_stream* stream, void* user_data, const void* input_buffer,
                              void* output_buffer, long num_frames);
@@ -117,10 +119,10 @@ CubebSink::CubebSink(std::string target_device_name) {
             LOG_WARNING(Audio_Sink, "Audio output device enumeration not supported");
         } else {
             const auto collection_end{collection.device + collection.count};
-            const auto device{std::find_if(collection.device, collection_end,
-                                           [&](const cubeb_device_info& device) {
-                                               return target_device_name == device.friendly_name;
-                                           })};
+            const auto device{
+                std::find_if(collection.device, collection_end, [&](const cubeb_device_info& info) {
+                    return target_device_name == info.friendly_name;
+                })};
             if (device != collection_end) {
                 output_device = device->devid;
             }
@@ -144,44 +146,59 @@ CubebSink::~CubebSink() {
 SinkStream& CubebSink::AcquireSinkStream(u32 sample_rate, u32 num_channels,
                                          const std::string& name) {
     sink_streams.push_back(
-        std::make_unique<SinkStreamImpl>(ctx, sample_rate, num_channels, output_device, name));
+        std::make_unique<CubebSinkStream>(ctx, sample_rate, num_channels, output_device, name));
     return *sink_streams.back();
 }
 
-long SinkStreamImpl::DataCallback(cubeb_stream* stream, void* user_data, const void* input_buffer,
-                                  void* output_buffer, long num_frames) {
-    SinkStreamImpl* impl = static_cast<SinkStreamImpl*>(user_data);
+long CubebSinkStream::DataCallback(cubeb_stream* stream, void* user_data, const void* input_buffer,
+                                   void* output_buffer, long num_frames) {
+    CubebSinkStream* impl = static_cast<CubebSinkStream*>(user_data);
     u8* buffer = reinterpret_cast<u8*>(output_buffer);
 
     if (!impl) {
         return {};
     }
 
-    std::lock_guard lock{impl->queue_mutex};
+    const std::size_t num_channels = impl->GetNumChannels();
+    const std::size_t samples_to_write = num_channels * num_frames;
+    std::size_t samples_written;
 
-    const size_t frames_to_write{
-        std::min(impl->queue.size() / impl->GetNumChannels(), static_cast<size_t>(num_frames))};
+    if (Settings::values.enable_audio_stretching) {
+        const std::vector<s16> in{impl->queue.Pop()};
+        const std::size_t num_in{in.size() / num_channels};
+        s16* const out{reinterpret_cast<s16*>(buffer)};
+        const std::size_t out_frames =
+            impl->time_stretch.Process(in.data(), num_in, out, num_frames);
+        samples_written = out_frames * num_channels;
 
-    memcpy(buffer, impl->queue.data(), frames_to_write * sizeof(s16) * impl->GetNumChannels());
-    impl->queue.erase(impl->queue.begin(),
-                      impl->queue.begin() + frames_to_write * impl->GetNumChannels());
+        if (impl->should_flush) {
+            impl->time_stretch.Flush();
+            impl->should_flush = false;
+        }
+    } else {
+        samples_written = impl->queue.Pop(buffer, samples_to_write);
+    }
 
-    if (frames_to_write < num_frames) {
-        // Fill the rest of the frames with silence
-        memset(buffer + frames_to_write * sizeof(s16) * impl->GetNumChannels(), 0,
-               (num_frames - frames_to_write) * sizeof(s16) * impl->GetNumChannels());
+    if (samples_written >= num_channels) {
+        std::memcpy(&impl->last_frame[0], buffer + (samples_written - num_channels) * sizeof(s16),
+                    num_channels * sizeof(s16));
+    }
+
+    // Fill the rest of the frames with last_frame
+    for (std::size_t i = samples_written; i < samples_to_write; i += num_channels) {
+        std::memcpy(buffer + i * sizeof(s16), &impl->last_frame[0], num_channels * sizeof(s16));
     }
 
     return num_frames;
 }
 
-void SinkStreamImpl::StateCallback(cubeb_stream* stream, void* user_data, cubeb_state state) {}
+void CubebSinkStream::StateCallback(cubeb_stream* stream, void* user_data, cubeb_state state) {}
 
 std::vector<std::string> ListCubebSinkDevices() {
     std::vector<std::string> device_list;
     cubeb* ctx;
 
-    if (cubeb_init(&ctx, "Citra Device Enumerator", nullptr) != CUBEB_OK) {
+    if (cubeb_init(&ctx, "yuzu Device Enumerator", nullptr) != CUBEB_OK) {
         LOG_CRITICAL(Audio_Sink, "cubeb_init failed");
         return {};
     }
@@ -190,7 +207,7 @@ std::vector<std::string> ListCubebSinkDevices() {
     if (cubeb_enumerate_devices(ctx, CUBEB_DEVICE_TYPE_OUTPUT, &collection) != CUBEB_OK) {
         LOG_WARNING(Audio_Sink, "Audio output device enumeration not supported");
     } else {
-        for (size_t i = 0; i < collection.count; i++) {
+        for (std::size_t i = 0; i < collection.count; i++) {
             const cubeb_device_info& device = collection.device[i];
             if (device.friendly_name) {
                 device_list.emplace_back(device.friendly_name);

src/audio_core/null_sink.h

@@ -21,6 +21,12 @@ public:
 private:
     struct NullSinkStreamImpl final : SinkStream {
         void EnqueueSamples(u32 /*num_channels*/, const std::vector<s16>& /*samples*/) override {}
+
+        std::size_t SamplesInQueue(u32 /*num_channels*/) const override {
+            return 0;
+        }
+
+        void Flush() override {}
     } null_sink_stream;
 };
 

src/audio_core/sink_details.cpp

@@ -24,7 +24,7 @@ const std::vector<SinkDetails> g_sink_details = {
                 [] { return std::vector<std::string>{"null"}; }},
 };
 
-const SinkDetails& GetSinkDetails(std::string sink_id) {
+const SinkDetails& GetSinkDetails(std::string_view sink_id) {
     auto iter =
         std::find_if(g_sink_details.begin(), g_sink_details.end(),
                      [sink_id](const auto& sink_detail) { return sink_detail.id == sink_id; });

src/audio_core/sink_details.h

@@ -6,6 +6,8 @@
 
 #include <functional>
 #include <memory>
+#include <string>
+#include <string_view>
 #include <utility>
 #include <vector>
 
@@ -30,6 +32,6 @@ struct SinkDetails {
 
 extern const std::vector<SinkDetails> g_sink_details;
 
-const SinkDetails& GetSinkDetails(std::string sink_id);
+const SinkDetails& GetSinkDetails(std::string_view sink_id);
 
 } // namespace AudioCore

src/audio_core/sink_stream.h

@@ -25,6 +25,10 @@ public:
      * @param samples Samples in interleaved stereo PCM16 format.
      */
     virtual void EnqueueSamples(u32 num_channels, const std::vector<s16>& samples) = 0;
+
+    virtual std::size_t SamplesInQueue(u32 num_channels) const = 0;
+
+    virtual void Flush() = 0;
 };
 
 using SinkStreamPtr = std::unique_ptr<SinkStream>;

src/audio_core/stream.cpp

@@ -7,16 +7,18 @@
 
 #include "audio_core/sink.h"
 #include "audio_core/sink_details.h"
+#include "audio_core/sink_stream.h"
 #include "audio_core/stream.h"
 #include "common/assert.h"
 #include "common/logging/log.h"
+#include "common/microprofile.h"
 #include "core/core_timing.h"
 #include "core/core_timing_util.h"
 #include "core/settings.h"
 
 namespace AudioCore {
 
-constexpr size_t MaxAudioBufferCount{32};
+constexpr std::size_t MaxAudioBufferCount{32};
 
 u32 Stream::GetNumChannels() const {
     switch (format) {
@@ -51,7 +53,7 @@ void Stream::Stop() {
 }
 
 s64 Stream::GetBufferReleaseCycles(const Buffer& buffer) const {
-    const size_t num_samples{buffer.GetSamples().size() / GetNumChannels()};
+    const std::size_t num_samples{buffer.GetSamples().size() / GetNumChannels()};
     return CoreTiming::usToCycles((static_cast<u64>(num_samples) * 1000000) / sample_rate);
 }
 
@@ -72,6 +74,7 @@ static void VolumeAdjustSamples(std::vector<s16>& samples) {
 void Stream::PlayNextBuffer() {
     if (!IsPlaying()) {
         // Ensure we are in playing state before playing the next buffer
+        sink_stream.Flush();
         return;
     }
 
@@ -82,6 +85,7 @@ void Stream::PlayNextBuffer() {
 
     if (queued_buffers.empty()) {
         // No queued buffers - we are effectively paused
+        sink_stream.Flush();
         return;
     }
 
@@ -89,12 +93,16 @@ void Stream::PlayNextBuffer() {
     queued_buffers.pop();
 
     VolumeAdjustSamples(active_buffer->Samples());
+
     sink_stream.EnqueueSamples(GetNumChannels(), active_buffer->GetSamples());
 
     CoreTiming::ScheduleEventThreadsafe(GetBufferReleaseCycles(*active_buffer), release_event, {});
 }
 
+MICROPROFILE_DEFINE(AudioOutput, "Audio", "ReleaseActiveBuffer", MP_RGB(100, 100, 255));
+
 void Stream::ReleaseActiveBuffer() {
+    MICROPROFILE_SCOPE(AudioOutput);
     ASSERT(active_buffer);
     released_buffers.push(std::move(active_buffer));
     release_callback();
@@ -115,9 +123,9 @@ bool Stream::ContainsBuffer(Buffer::Tag tag) const {
     return {};
 }
 
-std::vector<Buffer::Tag> Stream::GetTagsAndReleaseBuffers(size_t max_count) {
+std::vector<Buffer::Tag> Stream::GetTagsAndReleaseBuffers(std::size_t max_count) {
     std::vector<Buffer::Tag> tags;
-    for (size_t count = 0; count < max_count && !released_buffers.empty(); ++count) {
+    for (std::size_t count = 0; count < max_count && !released_buffers.empty(); ++count) {
         tags.push_back(released_buffers.front()->GetTag());
         released_buffers.pop();
     }

src/audio_core/stream.h

@@ -11,13 +11,16 @@
 #include <queue>
 
 #include "audio_core/buffer.h"
-#include "audio_core/sink_stream.h"
-#include "common/assert.h"
 #include "common/common_types.h"
-#include "core/core_timing.h"
+
+namespace CoreTiming {
+struct EventType;
+}
 
 namespace AudioCore {
 
+class SinkStream;
+
 /**
  * Represents an audio stream, which is a sequence of queued buffers, to be outputed by AudioOut
  */
@@ -49,7 +52,7 @@ public:
     bool ContainsBuffer(Buffer::Tag tag) const;
 
     /// Returns a vector of recently released buffers specified by tag
-    std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(size_t max_count);
+    std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(std::size_t max_count);
 
     /// Returns true if the stream is currently playing
     bool IsPlaying() const {
@@ -57,7 +60,7 @@ public:
     }
 
     /// Returns the number of queued buffers
-    size_t GetQueueSize() const {
+    std::size_t GetQueueSize() const {
         return queued_buffers.size();
     }
 

src/audio_core/time_stretch.cpp

@@ -0,0 +1,69 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <algorithm>
+#include <cmath>
+#include <cstddef>
+#include "audio_core/time_stretch.h"
+#include "common/logging/log.h"
+
+namespace AudioCore {
+
+TimeStretcher::TimeStretcher(u32 sample_rate, u32 channel_count)
+    : m_sample_rate(sample_rate), m_channel_count(channel_count) {
+    m_sound_touch.setChannels(channel_count);
+    m_sound_touch.setSampleRate(sample_rate);
+    m_sound_touch.setPitch(1.0);
+    m_sound_touch.setTempo(1.0);
+}
+
+void TimeStretcher::Clear() {
+    m_sound_touch.clear();
+}
+
+void TimeStretcher::Flush() {
+    m_sound_touch.flush();
+}
+
+std::size_t TimeStretcher::Process(const s16* in, std::size_t num_in, s16* out,
+                                   std::size_t num_out) {
+    const double time_delta = static_cast<double>(num_out) / m_sample_rate; // seconds
+
+    // We were given actual_samples number of samples, and num_samples were requested from us.
+    double current_ratio = static_cast<double>(num_in) / static_cast<double>(num_out);
+
+    const double max_latency = 1.0; // seconds
+    const double max_backlog = m_sample_rate * max_latency;
+    const double backlog_fullness = m_sound_touch.numSamples() / max_backlog;
+    if (backlog_fullness > 5.0) {
+        // Too many samples in backlog: Don't push anymore on
+        num_in = 0;
+    }
+
+    // We ideally want the backlog to be about 50% full.
+    // This gives some headroom both ways to prevent underflow and overflow.
+    // We tweak current_ratio to encourage this.
+    constexpr double tweak_time_scale = 0.05; // seconds
+    const double tweak_correction = (backlog_fullness - 0.5) * (time_delta / tweak_time_scale);
+    current_ratio *= std::pow(1.0 + 2.0 * tweak_correction, tweak_correction < 0 ? 3.0 : 1.0);
+
+    // This low-pass filter smoothes out variance in the calculated stretch ratio.
+    // The time-scale determines how responsive this filter is.
+    constexpr double lpf_time_scale = 2.0; // seconds
+    const double lpf_gain = 1.0 - std::exp(-time_delta / lpf_time_scale);
+    m_stretch_ratio += lpf_gain * (current_ratio - m_stretch_ratio);
+
+    // Place a lower limit of 5% speed.  When a game boots up, there will be
+    // many silence samples.  These do not need to be timestretched.
+    m_stretch_ratio = std::max(m_stretch_ratio, 0.05);
+    m_sound_touch.setTempo(m_stretch_ratio);
+
+    LOG_TRACE(Audio, "{:5}/{:5} ratio:{:0.6f} backlog:{:0.6f}", num_in, num_out, m_stretch_ratio,
+              backlog_fullness);
+
+    m_sound_touch.putSamples(in, static_cast<u32>(num_in));
+    return m_sound_touch.receiveSamples(out, static_cast<u32>(num_out));
+}
+
+} // namespace AudioCore

src/audio_core/time_stretch.h

@@ -0,0 +1,35 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <cstddef>
+#include <SoundTouch.h>
+#include "common/common_types.h"
+
+namespace AudioCore {
+
+class TimeStretcher {
+public:
+    TimeStretcher(u32 sample_rate, u32 channel_count);
+
+    /// @param in       Input sample buffer
+    /// @param num_in   Number of input frames in `in`
+    /// @param out      Output sample buffer
+    /// @param num_out  Desired number of output frames in `out`
+    /// @returns Actual number of frames written to `out`
+    std::size_t Process(const s16* in, std::size_t num_in, s16* out, std::size_t num_out);
+
+    void Clear();
+
+    void Flush();
+
+private:
+    u32 m_sample_rate;
+    u32 m_channel_count;
+    soundtouch::SoundTouch m_sound_touch;
+    double m_stretch_ratio = 1.0;
+};
+
+} // namespace AudioCore

src/common/CMakeLists.txt

@@ -1,13 +1,16 @@
 # Generate cpp with Git revision from template
-# Also if this is a CI build, add the build name (ie: Nightly, Bleeding Edge) to the scm_rev file as well
+# Also if this is a CI build, add the build name (ie: Nightly, Canary) to the scm_rev file as well
 set(REPO_NAME "")
+set(BUILD_VERSION "0")
 if ($ENV{CI})
   if ($ENV{TRAVIS})
     set(BUILD_REPOSITORY $ENV{TRAVIS_REPO_SLUG})
+    set(BUILD_TAG $ENV{TRAVIS_TAG})
   elseif($ENV{APPVEYOR})
     set(BUILD_REPOSITORY $ENV{APPVEYOR_REPO_NAME})
+    set(BUILD_TAG $ENV{APPVEYOR_REPO_TAG_NAME})
   endif()
-  # regex capture the string nightly or bleeding-edge into CMAKE_MATCH_1
+  # regex capture the string nightly or canary into CMAKE_MATCH_1
   string(REGEX MATCH "yuzu-emu/yuzu-?(.*)" OUTVAR ${BUILD_REPOSITORY})
   if (${CMAKE_MATCH_COUNT} GREATER 0)
     # capitalize the first letter of each word in the repo name.
@@ -16,10 +19,21 @@ if ($ENV{CI})
       string(SUBSTRING ${WORD} 0 1 FIRST_LETTER)
       string(SUBSTRING ${WORD} 1 -1 REMAINDER)
       string(TOUPPER ${FIRST_LETTER} FIRST_LETTER)
-      # this leaves a trailing space on the last word, but we actually want that
-      # because of how it's styled in the title bar.
-      set(REPO_NAME "${REPO_NAME}${FIRST_LETTER}${REMAINDER} ")
+      set(REPO_NAME "${REPO_NAME}${FIRST_LETTER}${REMAINDER}")
     endforeach()
+    if (BUILD_TAG)
+      string(REGEX MATCH "${CMAKE_MATCH_1}-([0-9]+)" OUTVAR ${BUILD_TAG})
+      if (${CMAKE_MATCH_COUNT} GREATER 0)
+        set(BUILD_VERSION ${CMAKE_MATCH_1})
+      endif()
+      if (BUILD_VERSION)
+        # This leaves a trailing space on the last word, but we actually want that
+        # because of how it's styled in the title bar.
+        set(BUILD_FULLNAME "${REPO_NAME} #${BUILD_VERSION} ")
+      else()
+        set(BUILD_FULLNAME "")
+      endif()
+    endif()
   endif()
 endif()
 configure_file("${CMAKE_CURRENT_SOURCE_DIR}/scm_rev.cpp.in" "${CMAKE_CURRENT_SOURCE_DIR}/scm_rev.cpp" @ONLY)
@@ -57,6 +71,7 @@ add_library(common STATIC
     param_package.cpp
     param_package.h
     quaternion.h
+    ring_buffer.h
     scm_rev.cpp
     scm_rev.h
     scope_exit.h

src/common/alignment.h

@@ -8,13 +8,13 @@
 namespace Common {
 
 template <typename T>
-constexpr T AlignUp(T value, size_t size) {
+constexpr T AlignUp(T value, std::size_t size) {
     static_assert(std::is_unsigned_v<T>, "T must be an unsigned value.");
     return static_cast<T>(value + (size - value % size) % size);
 }
 
 template <typename T>
-constexpr T AlignDown(T value, size_t size) {
+constexpr T AlignDown(T value, std::size_t size) {
     static_assert(std::is_unsigned_v<T>, "T must be an unsigned value.");
     return static_cast<T>(value - value % size);
 }

src/common/bit_field.h

@@ -129,8 +129,8 @@ private:
 
 public:
     /// Constants to allow limited introspection of fields if needed
-    static constexpr size_t position = Position;
-    static constexpr size_t bits = Bits;
+    static constexpr std::size_t position = Position;
+    static constexpr std::size_t bits = Bits;
     static constexpr StorageType mask = (((StorageTypeU)~0) >> (8 * sizeof(T) - bits)) << position;
 
     /**

src/common/bit_set.h

@@ -170,14 +170,14 @@ public:
             m_val |= (IntTy)1 << bit;
     }
 
-    static BitSet AllTrue(size_t count) {
+    static BitSet AllTrue(std::size_t count) {
         return BitSet(count == sizeof(IntTy) * 8 ? ~(IntTy)0 : (((IntTy)1 << count) - 1));
     }
 
-    Ref operator[](size_t bit) {
+    Ref operator[](std::size_t bit) {
         return Ref(this, (IntTy)1 << bit);
     }
-    const Ref operator[](size_t bit) const {
+    const Ref operator[](std::size_t bit) const {
         return (*const_cast<BitSet*>(this))[bit];
     }
     bool operator==(BitSet other) const {

src/common/cityhash.cpp

@@ -114,7 +114,7 @@ static uint64 HashLen16(uint64 u, uint64 v, uint64 mul) {
     return b;
 }
 
-static uint64 HashLen0to16(const char* s, size_t len) {
+static uint64 HashLen0to16(const char* s, std::size_t len) {
     if (len >= 8) {
         uint64 mul = k2 + len * 2;
         uint64 a = Fetch64(s) + k2;
@@ -141,7 +141,7 @@ static uint64 HashLen0to16(const char* s, size_t len) {
 
 // This probably works well for 16-byte strings as well, but it may be overkill
 // in that case.
-static uint64 HashLen17to32(const char* s, size_t len) {
+static uint64 HashLen17to32(const char* s, std::size_t len) {
     uint64 mul = k2 + len * 2;
     uint64 a = Fetch64(s) * k1;
     uint64 b = Fetch64(s + 8);
@@ -170,7 +170,7 @@ static pair<uint64, uint64> WeakHashLen32WithSeeds(const char* s, uint64 a, uint
 }
 
 // Return an 8-byte hash for 33 to 64 bytes.
-static uint64 HashLen33to64(const char* s, size_t len) {
+static uint64 HashLen33to64(const char* s, std::size_t len) {
     uint64 mul = k2 + len * 2;
     uint64 a = Fetch64(s) * k2;
     uint64 b = Fetch64(s + 8);
@@ -191,7 +191,7 @@ static uint64 HashLen33to64(const char* s, size_t len) {
     return b + x;
 }
 
-uint64 CityHash64(const char* s, size_t len) {
+uint64 CityHash64(const char* s, std::size_t len) {
     if (len <= 32) {
         if (len <= 16) {
             return HashLen0to16(s, len);
@@ -212,7 +212,7 @@ uint64 CityHash64(const char* s, size_t len) {
     x = x * k1 + Fetch64(s);
 
     // Decrease len to the nearest multiple of 64, and operate on 64-byte chunks.
-    len = (len - 1) & ~static_cast<size_t>(63);
+    len = (len - 1) & ~static_cast<std::size_t>(63);
     do {
         x = Rotate(x + y + v.first + Fetch64(s + 8), 37) * k1;
         y = Rotate(y + v.second + Fetch64(s + 48), 42) * k1;
@@ -229,17 +229,17 @@ uint64 CityHash64(const char* s, size_t len) {
                      HashLen16(v.second, w.second) + x);
 }
 
-uint64 CityHash64WithSeed(const char* s, size_t len, uint64 seed) {
+uint64 CityHash64WithSeed(const char* s, std::size_t len, uint64 seed) {
     return CityHash64WithSeeds(s, len, k2, seed);
 }
 
-uint64 CityHash64WithSeeds(const char* s, size_t len, uint64 seed0, uint64 seed1) {
+uint64 CityHash64WithSeeds(const char* s, std::size_t len, uint64 seed0, uint64 seed1) {
     return HashLen16(CityHash64(s, len) - seed0, seed1);
 }
 
 // A subroutine for CityHash128().  Returns a decent 128-bit hash for strings
 // of any length representable in signed long.  Based on City and Murmur.
-static uint128 CityMurmur(const char* s, size_t len, uint128 seed) {
+static uint128 CityMurmur(const char* s, std::size_t len, uint128 seed) {
     uint64 a = Uint128Low64(seed);
     uint64 b = Uint128High64(seed);
     uint64 c = 0;
@@ -269,7 +269,7 @@ static uint128 CityMurmur(const char* s, size_t len, uint128 seed) {
     return uint128(a ^ b, HashLen16(b, a));
 }
 
-uint128 CityHash128WithSeed(const char* s, size_t len, uint128 seed) {
+uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed) {
     if (len < 128) {
         return CityMurmur(s, len, seed);
     }
@@ -313,7 +313,7 @@ uint128 CityHash128WithSeed(const char* s, size_t len, uint128 seed) {
     w.first *= 9;
     v.first *= k0;
     // If 0 < len < 128, hash up to 4 chunks of 32 bytes each from the end of s.
-    for (size_t tail_done = 0; tail_done < len;) {
+    for (std::size_t tail_done = 0; tail_done < len;) {
         tail_done += 32;
         y = Rotate(x + y, 42) * k0 + v.second;
         w.first += Fetch64(s + len - tail_done + 16);
@@ -331,7 +331,7 @@ uint128 CityHash128WithSeed(const char* s, size_t len, uint128 seed) {
     return uint128(HashLen16(x + v.second, w.second) + y, HashLen16(x + w.second, y + v.second));
 }
 
-uint128 CityHash128(const char* s, size_t len) {
+uint128 CityHash128(const char* s, std::size_t len) {
     return len >= 16
                ? CityHash128WithSeed(s + 16, len - 16, uint128(Fetch64(s), Fetch64(s + 8) + k0))
                : CityHash128WithSeed(s, len, uint128(k0, k1));

src/common/cityhash.h

@@ -63,7 +63,7 @@
 
 #include <utility>
 #include <stdint.h>
-#include <stdlib.h> // for size_t.
+#include <stdlib.h> // for std::size_t.
 
 namespace Common {
 
@@ -77,22 +77,22 @@ inline uint64_t Uint128High64(const uint128& x) {
 }
 
 // Hash function for a byte array.
-uint64_t CityHash64(const char* buf, size_t len);
+uint64_t CityHash64(const char* buf, std::size_t len);
 
 // Hash function for a byte array.  For convenience, a 64-bit seed is also
 // hashed into the result.
-uint64_t CityHash64WithSeed(const char* buf, size_t len, uint64_t seed);
+uint64_t CityHash64WithSeed(const char* buf, std::size_t len, uint64_t seed);
 
 // Hash function for a byte array.  For convenience, two seeds are also
 // hashed into the result.
-uint64_t CityHash64WithSeeds(const char* buf, size_t len, uint64_t seed0, uint64_t seed1);
+uint64_t CityHash64WithSeeds(const char* buf, std::size_t len, uint64_t seed0, uint64_t seed1);
 
 // Hash function for a byte array.
-uint128 CityHash128(const char* s, size_t len);
+uint128 CityHash128(const char* s, std::size_t len);
 
 // Hash function for a byte array.  For convenience, a 128-bit seed is also
 // hashed into the result.
-uint128 CityHash128WithSeed(const char* s, size_t len, uint128 seed);
+uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed);
 
 // Hash 128 input bits down to 64 bits of output.
 // This is intended to be a reasonably good hash function.

src/common/file_util.cpp

@@ -76,7 +76,7 @@ namespace FileUtil {
 // Modifies argument.
 static void StripTailDirSlashes(std::string& fname) {
     if (fname.length() > 1) {
-        size_t i = fname.length();
+        std::size_t i = fname.length();
         while (i > 0 && fname[i - 1] == DIR_SEP_CHR)
             --i;
         fname.resize(i);
@@ -201,7 +201,7 @@ bool CreateFullPath(const std::string& fullPath) {
         return true;
     }
 
-    size_t position = 0;
+    std::size_t position = 0;
     while (true) {
         // Find next sub path
         position = fullPath.find(DIR_SEP_CHR, position);
@@ -299,7 +299,7 @@ bool Copy(const std::string& srcFilename, const std::string& destFilename) {
     std::array<char, 1024> buffer;
     while (!feof(input.get())) {
         // read input
-        size_t rnum = fread(buffer.data(), sizeof(char), buffer.size(), input.get());
+        std::size_t rnum = fread(buffer.data(), sizeof(char), buffer.size(), input.get());
         if (rnum != buffer.size()) {
             if (ferror(input.get()) != 0) {
                 LOG_ERROR(Common_Filesystem, "failed reading from source, {} --> {}: {}",
@@ -309,7 +309,7 @@ bool Copy(const std::string& srcFilename, const std::string& destFilename) {
         }
 
         // write output
-        size_t wnum = fwrite(buffer.data(), sizeof(char), rnum, output.get());
+        std::size_t wnum = fwrite(buffer.data(), sizeof(char), rnum, output.get());
         if (wnum != rnum) {
             LOG_ERROR(Common_Filesystem, "failed writing to output, {} --> {}: {}", srcFilename,
                       destFilename, GetLastErrorMsg());
@@ -756,15 +756,15 @@ std::string GetNANDRegistrationDir(bool system) {
     return GetUserPath(UserPath::NANDDir) + "user/Contents/registered/";
 }
 
-size_t WriteStringToFile(bool text_file, const std::string& str, const char* filename) {
+std::size_t WriteStringToFile(bool text_file, const std::string& str, const char* filename) {
     return FileUtil::IOFile(filename, text_file ? "w" : "wb").WriteBytes(str.data(), str.size());
 }
 
-size_t ReadFileToString(bool text_file, const char* filename, std::string& str) {
+std::size_t ReadFileToString(bool text_file, const char* filename, std::string& str) {
     IOFile file(filename, text_file ? "r" : "rb");
 
     if (!file.IsOpen())
-        return false;
+        return 0;
 
     str.resize(static_cast<u32>(file.GetSize()));
     return file.ReadArray(&str[0], str.size());
@@ -829,7 +829,7 @@ std::vector<std::string> SplitPathComponents(std::string_view filename) {
 std::string_view GetParentPath(std::string_view path) {
     const auto name_bck_index = path.rfind('\\');
     const auto name_fwd_index = path.rfind('/');
-    size_t name_index;
+    std::size_t name_index;
 
     if (name_bck_index == std::string_view::npos || name_fwd_index == std::string_view::npos) {
         name_index = std::min(name_bck_index, name_fwd_index);
@@ -868,7 +868,7 @@ std::string_view GetFilename(std::string_view path) {
 }
 
 std::string_view GetExtensionFromFilename(std::string_view name) {
-    const size_t index = name.rfind('.');
+    const std::size_t index = name.rfind('.');
 
     if (index == std::string_view::npos) {
         return {};

src/common/file_util.h

@@ -143,8 +143,9 @@ const std::string& GetExeDirectory();
 std::string AppDataRoamingDirectory();
 #endif
 
-size_t WriteStringToFile(bool text_file, const std::string& str, const char* filename);
-size_t ReadFileToString(bool text_file, const char* filename, std::string& str);
+std::size_t WriteStringToFile(bool text_file, const std::string& str, const char* filename);
+
+std::size_t ReadFileToString(bool text_file, const char* filename, std::string& str);
 
 /**
  * Splits the filename into 8.3 format
@@ -177,10 +178,10 @@ std::string_view RemoveTrailingSlash(std::string_view path);
 
 // Creates a new vector containing indices [first, last) from the original.
 template <typename T>
-std::vector<T> SliceVector(const std::vector<T>& vector, size_t first, size_t last) {
+std::vector<T> SliceVector(const std::vector<T>& vector, std::size_t first, std::size_t last) {
     if (first >= last)
         return {};
-    last = std::min<size_t>(last, vector.size());
+    last = std::min<std::size_t>(last, vector.size());
     return std::vector<T>(vector.begin() + first, vector.begin() + first + last);
 }
 
@@ -213,47 +214,47 @@ public:
     bool Close();
 
     template <typename T>
-    size_t ReadArray(T* data, size_t length) const {
+    std::size_t ReadArray(T* data, std::size_t length) const {
         static_assert(std::is_trivially_copyable_v<T>,
                       "Given array does not consist of trivially copyable objects");
 
         if (!IsOpen()) {
-            return std::numeric_limits<size_t>::max();
+            return std::numeric_limits<std::size_t>::max();
         }
 
         return std::fread(data, sizeof(T), length, m_file);
     }
 
     template <typename T>
-    size_t WriteArray(const T* data, size_t length) {
+    std::size_t WriteArray(const T* data, std::size_t length) {
         static_assert(std::is_trivially_copyable_v<T>,
                       "Given array does not consist of trivially copyable objects");
         if (!IsOpen()) {
-            return std::numeric_limits<size_t>::max();
+            return std::numeric_limits<std::size_t>::max();
         }
 
         return std::fwrite(data, sizeof(T), length, m_file);
     }
 
     template <typename T>
-    size_t ReadBytes(T* data, size_t length) const {
+    std::size_t ReadBytes(T* data, std::size_t length) const {
         static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable");
         return ReadArray(reinterpret_cast<char*>(data), length);
     }
 
     template <typename T>
-    size_t WriteBytes(const T* data, size_t length) {
+    std::size_t WriteBytes(const T* data, std::size_t length) {
         static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable");
         return WriteArray(reinterpret_cast<const char*>(data), length);
     }
 
     template <typename T>
-    size_t WriteObject(const T& object) {
+    std::size_t WriteObject(const T& object) {
         static_assert(!std::is_pointer_v<T>, "WriteObject arguments must not be a pointer");
         return WriteArray(&object, 1);
     }
 
-    size_t WriteString(const std::string& str) {
+    std::size_t WriteString(const std::string& str) {
         return WriteArray(str.c_str(), str.length());
     }
 

src/common/hash.h

@@ -17,7 +17,7 @@ namespace Common {
  * @param len Length of data (in bytes) to compute hash over
  * @returns 64-bit hash value that was computed over the data block
  */
-static inline u64 ComputeHash64(const void* data, size_t len) {
+static inline u64 ComputeHash64(const void* data, std::size_t len) {
     return CityHash64(static_cast<const char*>(data), len);
 }
 
@@ -63,7 +63,7 @@ struct HashableStruct {
         return !(*this == o);
     };
 
-    size_t Hash() const {
+    std::size_t Hash() const {
         return Common::ComputeStructHash64(state);
     }
 };

src/common/hex_util.cpp

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include "common/hex_util.h"
+#include "common/logging/log.h"
 
 namespace Common {
 
@@ -13,18 +14,29 @@ u8 ToHexNibble(char c1) {
         return c1 - 87;
     if (c1 >= 48 && c1 <= 57)
         return c1 - 48;
-    throw std::logic_error("Invalid hex digit");
+    LOG_ERROR(Common, "Invalid hex digit: 0x{:02X}", c1);
+    return 0;
 }
 
-std::array<u8, 16> operator""_array16(const char* str, size_t len) {
-    if (len != 32)
-        throw std::logic_error("Not of correct size.");
+std::array<u8, 16> operator""_array16(const char* str, std::size_t len) {
+    if (len != 32) {
+        LOG_ERROR(Common,
+                  "Attempting to parse string to array that is not of correct size (expected=32, "
+                  "actual={}).",
+                  len);
+        return {};
+    }
     return HexStringToArray<16>(str);
 }
 
-std::array<u8, 32> operator""_array32(const char* str, size_t len) {
-    if (len != 64)
-        throw std::logic_error("Not of correct size.");
+std::array<u8, 32> operator""_array32(const char* str, std::size_t len) {
+    if (len != 64) {
+        LOG_ERROR(Common,
+                  "Attempting to parse string to array that is not of correct size (expected=64, "
+                  "actual={}).",
+                  len);
+        return {};
+    }
     return HexStringToArray<32>(str);
 }
 

src/common/hex_util.h

@@ -14,20 +14,20 @@ namespace Common {
 
 u8 ToHexNibble(char c1);
 
-template <size_t Size, bool le = false>
+template <std::size_t Size, bool le = false>
 std::array<u8, Size> HexStringToArray(std::string_view str) {
     std::array<u8, Size> out{};
     if constexpr (le) {
-        for (size_t i = 2 * Size - 2; i <= 2 * Size; i -= 2)
+        for (std::size_t i = 2 * Size - 2; i <= 2 * Size; i -= 2)
             out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]);
     } else {
-        for (size_t i = 0; i < 2 * Size; i += 2)
+        for (std::size_t i = 0; i < 2 * Size; i += 2)
             out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]);
     }
     return out;
 }
 
-template <size_t Size>
+template <std::size_t Size>
 std::string HexArrayToString(std::array<u8, Size> array, bool upper = true) {
     std::string out;
     for (u8 c : array)
@@ -35,7 +35,7 @@ std::string HexArrayToString(std::array<u8, Size> array, bool upper = true) {
     return out;
 }
 
-std::array<u8, 0x10> operator"" _array16(const char* str, size_t len);
-std::array<u8, 0x20> operator"" _array32(const char* str, size_t len);
+std::array<u8, 0x10> operator"" _array16(const char* str, std::size_t len);
+std::array<u8, 0x20> operator"" _array32(const char* str, std::size_t len);
 
 } // namespace Common

src/common/logging/backend.cpp

@@ -135,7 +135,7 @@ FileBackend::FileBackend(const std::string& filename)
 void FileBackend::Write(const Entry& entry) {
     // prevent logs from going over the maximum size (in case its spamming and the user doesn't
     // know)
-    constexpr size_t MAX_BYTES_WRITTEN = 50 * 1024L * 1024L;
+    constexpr std::size_t MAX_BYTES_WRITTEN = 50 * 1024L * 1024L;
     if (!file.IsOpen() || bytes_written > MAX_BYTES_WRITTEN) {
         return;
     }

src/common/logging/backend.h

@@ -100,7 +100,7 @@ public:
 
 private:
     FileUtil::IOFile file;
-    size_t bytes_written;
+    std::size_t bytes_written;
 };
 
 void AddBackend(std::unique_ptr<Backend> backend);

src/common/logging/filter.cpp

@@ -71,7 +71,7 @@ void Filter::ResetAll(Level level) {
 }
 
 void Filter::SetClassLevel(Class log_class, Level level) {
-    class_levels[static_cast<size_t>(log_class)] = level;
+    class_levels[static_cast<std::size_t>(log_class)] = level;
 }
 
 void Filter::ParseFilterString(std::string_view filter_view) {
@@ -93,7 +93,8 @@ void Filter::ParseFilterString(std::string_view filter_view) {
 }
 
 bool Filter::CheckMessage(Class log_class, Level level) const {
-    return static_cast<u8>(level) >= static_cast<u8>(class_levels[static_cast<size_t>(log_class)]);
+    return static_cast<u8>(level) >=
+           static_cast<u8>(class_levels[static_cast<std::size_t>(log_class)]);
 }
 
 bool Filter::IsDebug() const {

src/common/logging/filter.h

@@ -19,7 +19,7 @@ namespace Log {
 class Filter {
 public:
     /// Initializes the filter with all classes having `default_level` as the minimum level.
-    Filter(Level default_level = Level::Info);
+    explicit Filter(Level default_level = Level::Info);
 
     /// Resets the filter so that all classes have `level` as the minimum displayed level.
     void ResetAll(Level level);
@@ -49,6 +49,6 @@ public:
     bool IsDebug() const;
 
 private:
-    std::array<Level, (size_t)Class::Count> class_levels;
+    std::array<Level, static_cast<std::size_t>(Class::Count)> class_levels;
 };
 } // namespace Log

src/common/logging/log.h

@@ -12,14 +12,14 @@ namespace Log {
 /// Specifies the severity or level of detail of the log message.
 enum class Level : u8 {
     Trace,    ///< Extremely detailed and repetitive debugging information that is likely to
-              ///  pollute logs.
+              ///< pollute logs.
     Debug,    ///< Less detailed debugging information.
     Info,     ///< Status information from important points during execution.
     Warning,  ///< Minor or potential problems found during execution of a task.
     Error,    ///< Major problems found during execution of a task that prevent it from being
-              ///  completed.
-    Critical, ///< Major problems during execution that threathen the stability of the entire
-              ///  application.
+              ///< completed.
+    Critical, ///< Major problems during execution that threaten the stability of the entire
+              ///< application.
 
     Count ///< Total number of logging levels
 };
@@ -49,7 +49,7 @@ enum class Class : ClassType {
     Kernel,            ///< The HLE implementation of the CTR kernel
     Kernel_SVC,        ///< Kernel system calls
     Service,           ///< HLE implementation of system services. Each major service
-                       ///  should have its own subclass.
+                       ///< should have its own subclass.
     Service_ACC,       ///< The ACC (Accounts) service
     Service_AM,        ///< The AM (Applet manager) service
     Service_AOC,       ///< The AOC (AddOn Content) service

src/common/logging/text_formatter.h

@@ -15,6 +15,6 @@ struct Entry;
 std::string FormatLogMessage(const Entry& entry);
 /// Formats and prints a log entry to stderr.
 void PrintMessage(const Entry& entry);
-/// Prints the same message as `PrintMessage`, but colored acoording to the severity level.
+/// Prints the same message as `PrintMessage`, but colored according to the severity level.
 void PrintColoredMessage(const Entry& entry);
 } // namespace Log

src/common/memory_util.cpp

@@ -25,7 +25,7 @@
 // This is purposely not a full wrapper for virtualalloc/mmap, but it
 // provides exactly the primitive operations that Dolphin needs.
 
-void* AllocateExecutableMemory(size_t size, bool low) {
+void* AllocateExecutableMemory(std::size_t size, bool low) {
 #if defined(_WIN32)
     void* ptr = VirtualAlloc(nullptr, size, MEM_COMMIT, PAGE_EXECUTE_READWRITE);
 #else
@@ -74,7 +74,7 @@ void* AllocateExecutableMemory(size_t size, bool low) {
     return ptr;
 }
 
-void* AllocateMemoryPages(size_t size) {
+void* AllocateMemoryPages(std::size_t size) {
 #ifdef _WIN32
     void* ptr = VirtualAlloc(nullptr, size, MEM_COMMIT, PAGE_READWRITE);
 #else
@@ -90,7 +90,7 @@ void* AllocateMemoryPages(size_t size) {
     return ptr;
 }
 
-void* AllocateAlignedMemory(size_t size, size_t alignment) {
+void* AllocateAlignedMemory(std::size_t size, std::size_t alignment) {
 #ifdef _WIN32
     void* ptr = _aligned_malloc(size, alignment);
 #else
@@ -109,7 +109,7 @@ void* AllocateAlignedMemory(size_t size, size_t alignment) {
     return ptr;
 }
 
-void FreeMemoryPages(void* ptr, size_t size) {
+void FreeMemoryPages(void* ptr, std::size_t size) {
     if (ptr) {
 #ifdef _WIN32
         if (!VirtualFree(ptr, 0, MEM_RELEASE))
@@ -130,7 +130,7 @@ void FreeAlignedMemory(void* ptr) {
     }
 }
 
-void WriteProtectMemory(void* ptr, size_t size, bool allowExecute) {
+void WriteProtectMemory(void* ptr, std::size_t size, bool allowExecute) {
 #ifdef _WIN32
     DWORD oldValue;
     if (!VirtualProtect(ptr, size, allowExecute ? PAGE_EXECUTE_READ : PAGE_READONLY, &oldValue))
@@ -140,7 +140,7 @@ void WriteProtectMemory(void* ptr, size_t size, bool allowExecute) {
 #endif
 }
 
-void UnWriteProtectMemory(void* ptr, size_t size, bool allowExecute) {
+void UnWriteProtectMemory(void* ptr, std::size_t size, bool allowExecute) {
 #ifdef _WIN32
     DWORD oldValue;
     if (!VirtualProtect(ptr, size, allowExecute ? PAGE_EXECUTE_READWRITE : PAGE_READWRITE,

src/common/memory_util.h

@@ -7,13 +7,13 @@
 #include <cstddef>
 #include <string>
 
-void* AllocateExecutableMemory(size_t size, bool low = true);
-void* AllocateMemoryPages(size_t size);
-void FreeMemoryPages(void* ptr, size_t size);
-void* AllocateAlignedMemory(size_t size, size_t alignment);
+void* AllocateExecutableMemory(std::size_t size, bool low = true);
+void* AllocateMemoryPages(std::size_t size);
+void FreeMemoryPages(void* ptr, std::size_t size);
+void* AllocateAlignedMemory(std::size_t size, std::size_t alignment);
 void FreeAlignedMemory(void* ptr);
-void WriteProtectMemory(void* ptr, size_t size, bool executable = false);
-void UnWriteProtectMemory(void* ptr, size_t size, bool allowExecute = false);
+void WriteProtectMemory(void* ptr, std::size_t size, bool executable = false);
+void UnWriteProtectMemory(void* ptr, std::size_t size, bool allowExecute = false);
 std::string MemUsage();
 
 inline int GetPageSize() {

src/common/misc.cpp

@@ -16,7 +16,7 @@
 // Call directly after the command or use the error num.
 // This function might change the error code.
 std::string GetLastErrorMsg() {
-    static const size_t buff_size = 255;
+    static const std::size_t buff_size = 255;
     char err_str[buff_size];
 
 #ifdef _WIN32

src/common/ring_buffer.h

@@ -0,0 +1,119 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <algorithm>
+#include <array>
+#include <atomic>
+#include <cstddef>
+#include <cstring>
+#include <new>
+#include <type_traits>
+#include <vector>
+#include "common/common_types.h"
+
+namespace Common {
+
+/// SPSC ring buffer
+/// @tparam T            Element type
+/// @tparam capacity     Number of slots in ring buffer
+/// @tparam granularity  Slot size in terms of number of elements
+template <typename T, std::size_t capacity, std::size_t granularity = 1>
+class RingBuffer {
+    /// A "slot" is made of `granularity` elements of `T`.
+    static constexpr std::size_t slot_size = granularity * sizeof(T);
+    // T must be safely memcpy-able and have a trivial default constructor.
+    static_assert(std::is_trivial_v<T>);
+    // Ensure capacity is sensible.
+    static_assert(capacity < std::numeric_limits<std::size_t>::max() / 2 / granularity);
+    static_assert((capacity & (capacity - 1)) == 0, "capacity must be a power of two");
+    // Ensure lock-free.
+    static_assert(std::atomic_size_t::is_always_lock_free);
+
+public:
+    /// Pushes slots into the ring buffer
+    /// @param new_slots   Pointer to the slots to push
+    /// @param slot_count  Number of slots to push
+    /// @returns The number of slots actually pushed
+    std::size_t Push(const void* new_slots, std::size_t slot_count) {
+        const std::size_t write_index = m_write_index.load();
+        const std::size_t slots_free = capacity + m_read_index.load() - write_index;
+        const std::size_t push_count = std::min(slot_count, slots_free);
+
+        const std::size_t pos = write_index % capacity;
+        const std::size_t first_copy = std::min(capacity - pos, push_count);
+        const std::size_t second_copy = push_count - first_copy;
+
+        const char* in = static_cast<const char*>(new_slots);
+        std::memcpy(m_data.data() + pos * granularity, in, first_copy * slot_size);
+        in += first_copy * slot_size;
+        std::memcpy(m_data.data(), in, second_copy * slot_size);
+
+        m_write_index.store(write_index + push_count);
+
+        return push_count;
+    }
+
+    std::size_t Push(const std::vector<T>& input) {
+        return Push(input.data(), input.size());
+    }
+
+    /// Pops slots from the ring buffer
+    /// @param output     Where to store the popped slots
+    /// @param max_slots  Maximum number of slots to pop
+    /// @returns The number of slots actually popped
+    std::size_t Pop(void* output, std::size_t max_slots = ~std::size_t(0)) {
+        const std::size_t read_index = m_read_index.load();
+        const std::size_t slots_filled = m_write_index.load() - read_index;
+        const std::size_t pop_count = std::min(slots_filled, max_slots);
+
+        const std::size_t pos = read_index % capacity;
+        const std::size_t first_copy = std::min(capacity - pos, pop_count);
+        const std::size_t second_copy = pop_count - first_copy;
+
+        char* out = static_cast<char*>(output);
+        std::memcpy(out, m_data.data() + pos * granularity, first_copy * slot_size);
+        out += first_copy * slot_size;
+        std::memcpy(out, m_data.data(), second_copy * slot_size);
+
+        m_read_index.store(read_index + pop_count);
+
+        return pop_count;
+    }
+
+    std::vector<T> Pop(std::size_t max_slots = ~std::size_t(0)) {
+        std::vector<T> out(std::min(max_slots, capacity) * granularity);
+        const std::size_t count = Pop(out.data(), out.size() / granularity);
+        out.resize(count * granularity);
+        return out;
+    }
+
+    /// @returns Number of slots used
+    std::size_t Size() const {
+        return m_write_index.load() - m_read_index.load();
+    }
+
+    /// @returns Maximum size of ring buffer
+    constexpr std::size_t Capacity() const {
+        return capacity;
+    }
+
+private:
+    // It is important to align the below variables for performance reasons:
+    // Having them on the same cache-line would result in false-sharing between them.
+    // TODO: Remove this ifdef whenever clang and GCC support
+    //       std::hardware_destructive_interference_size.
+#if defined(_MSC_VER) && _MSC_VER >= 1911
+    alignas(std::hardware_destructive_interference_size) std::atomic_size_t m_read_index{0};
+    alignas(std::hardware_destructive_interference_size) std::atomic_size_t m_write_index{0};
+#else
+    alignas(128) std::atomic_size_t m_read_index{0};
+    alignas(128) std::atomic_size_t m_write_index{0};
+#endif
+
+    std::array<T, granularity * capacity> m_data;
+};
+
+} // namespace Common

src/common/scm_rev.cpp.in

@@ -9,6 +9,8 @@
 #define GIT_DESC     "@GIT_DESC@"
 #define BUILD_NAME   "@REPO_NAME@"
 #define BUILD_DATE   "@BUILD_DATE@"
+#define BUILD_FULLNAME "@BUILD_FULLNAME@"
+#define BUILD_VERSION "@BUILD_VERSION@"
 
 namespace Common {
 
@@ -17,6 +19,8 @@ const char g_scm_branch[]   = GIT_BRANCH;
 const char g_scm_desc[]     = GIT_DESC;
 const char g_build_name[]   = BUILD_NAME;
 const char g_build_date[]   = BUILD_DATE;
+const char g_build_fullname[] = BUILD_FULLNAME;
+const char g_build_version[]  = BUILD_VERSION;
 
 } // namespace
 

src/common/scm_rev.h

@@ -11,5 +11,7 @@ extern const char g_scm_branch[];
 extern const char g_scm_desc[];
 extern const char g_build_name[];
 extern const char g_build_date[];
+extern const char g_build_fullname[];
+extern const char g_build_version[];
 
 } // namespace Common

src/common/string_util.cpp

@@ -37,7 +37,7 @@ std::string ToUpper(std::string str) {
 }
 
 // For Debugging. Read out an u8 array.
-std::string ArrayToString(const u8* data, size_t size, int line_len, bool spaces) {
+std::string ArrayToString(const u8* data, std::size_t size, int line_len, bool spaces) {
     std::ostringstream oss;
     oss << std::setfill('0') << std::hex;
 
@@ -60,7 +60,7 @@ std::string StringFromBuffer(const std::vector<u8>& data) {
 
 // Turns "  hej " into "hej". Also handles tabs.
 std::string StripSpaces(const std::string& str) {
-    const size_t s = str.find_first_not_of(" \t\r\n");
+    const std::size_t s = str.find_first_not_of(" \t\r\n");
 
     if (str.npos != s)
         return str.substr(s, str.find_last_not_of(" \t\r\n") - s + 1);
@@ -121,10 +121,10 @@ bool SplitPath(const std::string& full_path, std::string* _pPath, std::string* _
     if (full_path.empty())
         return false;
 
-    size_t dir_end = full_path.find_last_of("/"
+    std::size_t dir_end = full_path.find_last_of("/"
 // windows needs the : included for something like just "C:" to be considered a directory
 #ifdef _WIN32
-                                            "\\:"
+                                                 "\\:"
 #endif
     );
     if (std::string::npos == dir_end)
@@ -132,7 +132,7 @@ bool SplitPath(const std::string& full_path, std::string* _pPath, std::string* _
     else
         dir_end += 1;
 
-    size_t fname_end = full_path.rfind('.');
+    std::size_t fname_end = full_path.rfind('.');
     if (fname_end < dir_end || std::string::npos == fname_end)
         fname_end = full_path.size();
 
@@ -172,7 +172,7 @@ void SplitString(const std::string& str, const char delim, std::vector<std::stri
 }
 
 std::string TabsToSpaces(int tab_size, std::string in) {
-    size_t i = 0;
+    std::size_t i = 0;
 
     while ((i = in.find('\t')) != std::string::npos) {
         in.replace(i, 1, tab_size, ' ');
@@ -182,7 +182,7 @@ std::string TabsToSpaces(int tab_size, std::string in) {
 }
 
 std::string ReplaceAll(std::string result, const std::string& src, const std::string& dest) {
-    size_t pos = 0;
+    std::size_t pos = 0;
 
     if (src == dest)
         return result;
@@ -280,22 +280,22 @@ static std::string CodeToUTF8(const char* fromcode, const std::basic_string<T>&
         return {};
     }
 
-    const size_t in_bytes = sizeof(T) * input.size();
+    const std::size_t in_bytes = sizeof(T) * input.size();
     // Multiply by 4, which is the max number of bytes to encode a codepoint
-    const size_t out_buffer_size = 4 * in_bytes;
+    const std::size_t out_buffer_size = 4 * in_bytes;
 
     std::string out_buffer(out_buffer_size, '\0');
 
     auto src_buffer = &input[0];
-    size_t src_bytes = in_bytes;
+    std::size_t src_bytes = in_bytes;
     auto dst_buffer = &out_buffer[0];
-    size_t dst_bytes = out_buffer.size();
+    std::size_t dst_bytes = out_buffer.size();
 
     while (0 != src_bytes) {
-        size_t const iconv_result =
+        std::size_t const iconv_result =
             iconv(conv_desc, (char**)(&src_buffer), &src_bytes, &dst_buffer, &dst_bytes);
 
-        if (static_cast<size_t>(-1) == iconv_result) {
+        if (static_cast<std::size_t>(-1) == iconv_result) {
             if (EILSEQ == errno || EINVAL == errno) {
                 // Try to skip the bad character
                 if (0 != src_bytes) {
@@ -326,22 +326,22 @@ std::u16string UTF8ToUTF16(const std::string& input) {
         return {};
     }
 
-    const size_t in_bytes = sizeof(char) * input.size();
+    const std::size_t in_bytes = sizeof(char) * input.size();
     // Multiply by 4, which is the max number of bytes to encode a codepoint
-    const size_t out_buffer_size = 4 * sizeof(char16_t) * in_bytes;
+    const std::size_t out_buffer_size = 4 * sizeof(char16_t) * in_bytes;
 
     std::u16string out_buffer(out_buffer_size, char16_t{});
 
     char* src_buffer = const_cast<char*>(&input[0]);
-    size_t src_bytes = in_bytes;
+    std::size_t src_bytes = in_bytes;
     char* dst_buffer = (char*)(&out_buffer[0]);
-    size_t dst_bytes = out_buffer.size();
+    std::size_t dst_bytes = out_buffer.size();
 
     while (0 != src_bytes) {
-        size_t const iconv_result =
+        std::size_t const iconv_result =
             iconv(conv_desc, &src_buffer, &src_bytes, &dst_buffer, &dst_bytes);
 
-        if (static_cast<size_t>(-1) == iconv_result) {
+        if (static_cast<std::size_t>(-1) == iconv_result) {
             if (EILSEQ == errno || EINVAL == errno) {
                 // Try to skip the bad character
                 if (0 != src_bytes) {
@@ -381,8 +381,8 @@ std::string SHIFTJISToUTF8(const std::string& input) {
 
 #endif
 
-std::string StringFromFixedZeroTerminatedBuffer(const char* buffer, size_t max_len) {
-    size_t len = 0;
+std::string StringFromFixedZeroTerminatedBuffer(const char* buffer, std::size_t max_len) {
+    std::size_t len = 0;
     while (len < max_len && buffer[len] != '\0')
         ++len;
 

src/common/string_util.h

@@ -19,7 +19,7 @@ std::string ToLower(std::string str);
 /// Make a string uppercase
 std::string ToUpper(std::string str);
 
-std::string ArrayToString(const u8* data, size_t size, int line_len = 20, bool spaces = true);
+std::string ArrayToString(const u8* data, std::size_t size, int line_len = 20, bool spaces = true);
 
 std::string StringFromBuffer(const std::vector<u8>& data);
 
@@ -118,7 +118,7 @@ bool ComparePartialString(InIt begin, InIt end, const char* other) {
  * Creates a std::string from a fixed-size NUL-terminated char buffer. If the buffer isn't
  * NUL-terminated then the string ends at max_len characters.
  */
-std::string StringFromFixedZeroTerminatedBuffer(const char* buffer, size_t max_len);
+std::string StringFromFixedZeroTerminatedBuffer(const char* buffer, std::size_t max_len);
 
 /**
  * Attempts to trim an arbitrary prefix from `path`, leaving only the part starting at `root`. It's

src/common/thread.h

@@ -60,12 +60,12 @@ private:
 
 class Barrier {
 public:
-    explicit Barrier(size_t count_) : count(count_), waiting(0), generation(0) {}
+    explicit Barrier(std::size_t count_) : count(count_), waiting(0), generation(0) {}
 
     /// Blocks until all "count" threads have called Sync()
     void Sync() {
         std::unique_lock<std::mutex> lk(mutex);
-        const size_t current_generation = generation;
+        const std::size_t current_generation = generation;
 
         if (++waiting == count) {
             generation++;
@@ -80,21 +80,13 @@ public:
 private:
     std::condition_variable condvar;
     std::mutex mutex;
-    const size_t count;
-    size_t waiting;
-    size_t generation; // Incremented once each time the barrier is used
+    const std::size_t count;
+    std::size_t waiting;
+    std::size_t generation; // Incremented once each time the barrier is used
 };
 
 void SleepCurrentThread(int ms);
 void SwitchCurrentThread(); // On Linux, this is equal to sleep 1ms
-
-// Use this function during a spin-wait to make the current thread
-// relax while another thread is working. This may be more efficient
-// than using events because event functions use kernel calls.
-inline void YieldCPU() {
-    std::this_thread::yield();
-}
-
 void SetCurrentThreadName(const char* name);
 
 } // namespace Common

src/common/x64/xbyak_abi.h

@@ -97,7 +97,7 @@ const BitSet32 ABI_ALL_CALLEE_SAVED = BuildRegSet({
     Xbyak::util::xmm15,
 });
 
-constexpr size_t ABI_SHADOW_SPACE = 0x20;
+constexpr std::size_t ABI_SHADOW_SPACE = 0x20;
 
 #else
 
@@ -147,22 +147,23 @@ const BitSet32 ABI_ALL_CALLEE_SAVED = BuildRegSet({
     Xbyak::util::r15,
 });
 
-constexpr size_t ABI_SHADOW_SPACE = 0;
+constexpr std::size_t ABI_SHADOW_SPACE = 0;
 
 #endif
 
-inline void ABI_CalculateFrameSize(BitSet32 regs, size_t rsp_alignment, size_t needed_frame_size,
-                                   s32* out_subtraction, s32* out_xmm_offset) {
+inline void ABI_CalculateFrameSize(BitSet32 regs, std::size_t rsp_alignment,
+                                   std::size_t needed_frame_size, s32* out_subtraction,
+                                   s32* out_xmm_offset) {
     int count = (regs & ABI_ALL_GPRS).Count();
     rsp_alignment -= count * 8;
-    size_t subtraction = 0;
+    std::size_t subtraction = 0;
     int xmm_count = (regs & ABI_ALL_XMMS).Count();
     if (xmm_count) {
         // If we have any XMMs to save, we must align the stack here.
         subtraction = rsp_alignment & 0xF;
     }
     subtraction += 0x10 * xmm_count;
-    size_t xmm_base_subtraction = subtraction;
+    std::size_t xmm_base_subtraction = subtraction;
     subtraction += needed_frame_size;
     subtraction += ABI_SHADOW_SPACE;
     // Final alignment.
@@ -173,8 +174,9 @@ inline void ABI_CalculateFrameSize(BitSet32 regs, size_t rsp_alignment, size_t n
     *out_xmm_offset = (s32)(subtraction - xmm_base_subtraction);
 }
 
-inline size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs,
-                                              size_t rsp_alignment, size_t needed_frame_size = 0) {
+inline std::size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs,
+                                                   std::size_t rsp_alignment,
+                                                   std::size_t needed_frame_size = 0) {
     s32 subtraction, xmm_offset;
     ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size, &subtraction, &xmm_offset);
 
@@ -195,7 +197,8 @@ inline size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet
 }
 
 inline void ABI_PopRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs,
-                                           size_t rsp_alignment, size_t needed_frame_size = 0) {
+                                           std::size_t rsp_alignment,
+                                           std::size_t needed_frame_size = 0) {
     s32 subtraction, xmm_offset;
     ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size, &subtraction, &xmm_offset);
 

src/common/x64/xbyak_util.h

@@ -34,7 +34,7 @@ inline bool IsWithin2G(const Xbyak::CodeGenerator& code, uintptr_t target) {
 template <typename T>
 inline void CallFarFunction(Xbyak::CodeGenerator& code, const T f) {
     static_assert(std::is_pointer_v<T>, "Argument must be a (function) pointer.");
-    size_t addr = reinterpret_cast<size_t>(f);
+    std::size_t addr = reinterpret_cast<std::size_t>(f);
     if (IsWithin2G(code, addr)) {
         code.call(f);
     } else {

src/core/CMakeLists.txt

@@ -20,6 +20,8 @@ add_library(core STATIC
     crypto/key_manager.h
     crypto/ctr_encryption_layer.cpp
     crypto/ctr_encryption_layer.h
+    crypto/xts_encryption_layer.cpp
+    crypto/xts_encryption_layer.h
     file_sys/bis_factory.cpp
     file_sys/bis_factory.h
     file_sys/card_image.cpp
@@ -33,8 +35,12 @@ add_library(core STATIC
     file_sys/mode.h
     file_sys/nca_metadata.cpp
     file_sys/nca_metadata.h
+    file_sys/nca_patch.cpp
+    file_sys/nca_patch.h
     file_sys/partition_filesystem.cpp
     file_sys/partition_filesystem.h
+    file_sys/patch_manager.cpp
+    file_sys/patch_manager.h
     file_sys/program_metadata.cpp
     file_sys/program_metadata.h
     file_sys/registered_cache.cpp
@@ -47,6 +53,8 @@ add_library(core STATIC
     file_sys/savedata_factory.h
     file_sys/sdmc_factory.cpp
     file_sys/sdmc_factory.h
+    file_sys/submission_package.cpp
+    file_sys/submission_package.h
     file_sys/vfs.cpp
     file_sys/vfs.h
     file_sys/vfs_concat.cpp
@@ -57,6 +65,8 @@ add_library(core STATIC
     file_sys/vfs_real.h
     file_sys/vfs_vector.cpp
     file_sys/vfs_vector.h
+    file_sys/xts_archive.cpp
+    file_sys/xts_archive.h
     frontend/emu_window.cpp
     frontend/emu_window.h
     frontend/framebuffer_layout.cpp
@@ -347,12 +357,16 @@ add_library(core STATIC
     loader/linker.h
     loader/loader.cpp
     loader/loader.h
+    loader/nax.cpp
+    loader/nax.h
     loader/nca.cpp
     loader/nca.h
     loader/nro.cpp
     loader/nro.h
     loader/nso.cpp
     loader/nso.h
+    loader/nsp.cpp
+    loader/nsp.h
     loader/xci.cpp
     loader/xci.h
     memory.cpp
@@ -374,7 +388,7 @@ add_library(core STATIC
 create_target_directory_groups(core)
 
 target_link_libraries(core PUBLIC common PRIVATE audio_core video_core)
-target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt lz4_static mbedtls opus unicorn)
+target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt lz4_static mbedtls opus unicorn open_source_archives)
 
 if (ARCHITECTURE_x86_64)
     target_sources(core PRIVATE

src/core/arm/arm_interface.h

@@ -6,9 +6,14 @@
 
 #include <array>
 #include "common/common_types.h"
-#include "core/hle/kernel/vm_manager.h"
 
-/// Generic ARM11 CPU interface
+namespace Kernel {
+enum class VMAPermission : u8;
+}
+
+namespace Core {
+
+/// Generic ARMv8 CPU interface
 class ARM_Interface : NonCopyable {
 public:
     virtual ~ARM_Interface() {}
@@ -17,9 +22,9 @@ public:
         std::array<u64, 31> cpu_registers;
         u64 sp;
         u64 pc;
-        u64 cpsr;
-        std::array<u128, 32> fpu_registers;
-        u64 fpscr;
+        u64 pstate;
+        std::array<u128, 32> vector_registers;
+        u64 fpcr;
     };
 
     /// Runs the CPU until an event happens
@@ -29,11 +34,11 @@ public:
     virtual void Step() = 0;
 
     /// Maps a backing memory region for the CPU
-    virtual void MapBackingMemory(VAddr address, size_t size, u8* memory,
+    virtual void MapBackingMemory(VAddr address, std::size_t size, u8* memory,
                                   Kernel::VMAPermission perms) = 0;
 
     /// Unmaps a region of memory that was previously mapped using MapBackingMemory
-    virtual void UnmapMemory(VAddr address, size_t size) = 0;
+    virtual void UnmapMemory(VAddr address, std::size_t size) = 0;
 
     /// Clear all instruction cache
     virtual void ClearInstructionCache() = 0;
@@ -67,42 +72,50 @@ public:
      */
     virtual void SetReg(int index, u64 value) = 0;
 
-    virtual u128 GetExtReg(int index) const = 0;
-
-    virtual void SetExtReg(int index, u128 value) = 0;
+    /**
+     * Gets the value of a specified vector register.
+     *
+     * @param index The index of the vector register.
+     * @return the value within the vector register.
+     */
+    virtual u128 GetVectorReg(int index) const = 0;
 
     /**
-     * Gets the value of a VFP register
-     * @param index Register index (0-31)
-     * @return Returns the value in the register
+     * Sets a given value into a vector register.
+     *
+     * @param index The index of the vector register.
+     * @param value The new value to place in the register.
      */
-    virtual u32 GetVFPReg(int index) const = 0;
+    virtual void SetVectorReg(int index, u128 value) = 0;
 
     /**
-     * Sets a VFP register to the given value
-     * @param index Register index (0-31)
-     * @param value Value to set register to
+     * Get the current PSTATE register
+     * @return Returns the value of the PSTATE register
      */
-    virtual void SetVFPReg(int index, u32 value) = 0;
+    virtual u32 GetPSTATE() const = 0;
 
     /**
-     * Get the current CPSR register
-     * @return Returns the value of the CPSR register
+     * Set the current PSTATE register
+     * @param pstate Value to set PSTATE to
      */
-    virtual u32 GetCPSR() const = 0;
-
-    /**
-     * Set the current CPSR register
-     * @param cpsr Value to set CPSR to
-     */
-    virtual void SetCPSR(u32 cpsr) = 0;
+    virtual void SetPSTATE(u32 pstate) = 0;
 
     virtual VAddr GetTlsAddress() const = 0;
 
     virtual void SetTlsAddress(VAddr address) = 0;
 
+    /**
+     * Gets the value within the TPIDR_EL0 (read/write software thread ID) register.
+     *
+     * @return the value within the register.
+     */
     virtual u64 GetTPIDR_EL0() const = 0;
 
+    /**
+     * Sets a new value within the TPIDR_EL0 (read/write software thread ID) register.
+     *
+     * @param value The new value to place in the register.
+     */
     virtual void SetTPIDR_EL0(u64 value) = 0;
 
     /**
@@ -117,8 +130,11 @@ public:
      */
     virtual void LoadContext(const ThreadContext& ctx) = 0;
 
+    /// Clears the exclusive monitor's state.
     virtual void ClearExclusiveState() = 0;
 
     /// Prepare core for thread reschedule (if needed to correctly handle state)
     virtual void PrepareReschedule() = 0;
 };
+
+} // namespace Core

src/core/arm/dynarmic/arm_dynarmic.cpp

@@ -7,13 +7,19 @@
 #include <dynarmic/A64/a64.h>
 #include <dynarmic/A64/config.h>
 #include "common/logging/log.h"
+#include "common/microprofile.h"
 #include "core/arm/dynarmic/arm_dynarmic.h"
 #include "core/core.h"
+#include "core/core_cpu.h"
 #include "core/core_timing.h"
+#include "core/gdbstub/gdbstub.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/svc.h"
+#include "core/hle/kernel/vm_manager.h"
 #include "core/memory.h"
 
+namespace Core {
+
 using Vector = Dynarmic::A64::Vector;
 
 class ARM_Dynarmic_Callbacks : public Dynarmic::A64::UserCallbacks {
@@ -54,7 +60,7 @@ public:
         Memory::Write64(vaddr + 8, value[1]);
     }
 
-    void InterpreterFallback(u64 pc, size_t num_instructions) override {
+    void InterpreterFallback(u64 pc, std::size_t num_instructions) override {
         LOG_INFO(Core_ARM, "Unicorn fallback @ 0x{:X} for {} instructions (instr = {:08X})", pc,
                  num_instructions, MemoryReadCode(pc));
 
@@ -75,9 +81,20 @@ public:
         case Dynarmic::A64::Exception::SendEventLocal:
         case Dynarmic::A64::Exception::Yield:
             return;
+        case Dynarmic::A64::Exception::Breakpoint:
+            if (GDBStub::IsServerEnabled()) {
+                parent.jit->HaltExecution();
+                parent.SetPC(pc);
+                Kernel::Thread* thread = Kernel::GetCurrentThread();
+                parent.SaveContext(thread->context);
+                GDBStub::Break();
+                GDBStub::SendTrap(thread, 5);
+                return;
+            }
+            [[fallthrough]];
         default:
             ASSERT_MSG(false, "ExceptionRaised(exception = {}, pc = {:X})",
-                       static_cast<size_t>(exception), pc);
+                       static_cast<std::size_t>(exception), pc);
         }
     }
 
@@ -106,7 +123,7 @@ public:
     }
 
     ARM_Dynarmic& parent;
-    size_t num_interpreted_instructions = 0;
+    std::size_t num_interpreted_instructions = 0;
     u64 tpidrro_el0 = 0;
     u64 tpidr_el0 = 0;
 };
@@ -140,7 +157,10 @@ std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const {
     return std::make_unique<Dynarmic::A64::Jit>(config);
 }
 
+MICROPROFILE_DEFINE(ARM_Jit_Dynarmic, "ARM JIT", "Dynarmic", MP_RGB(255, 64, 64));
+
 void ARM_Dynarmic::Run() {
+    MICROPROFILE_SCOPE(ARM_Jit_Dynarmic);
     ASSERT(Memory::GetCurrentPageTable() == current_page_table);
 
     jit->Run();
@@ -150,7 +170,8 @@ void ARM_Dynarmic::Step() {
     cb->InterpreterFallback(jit->GetPC(), 1);
 }
 
-ARM_Dynarmic::ARM_Dynarmic(std::shared_ptr<ExclusiveMonitor> exclusive_monitor, size_t core_index)
+ARM_Dynarmic::ARM_Dynarmic(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
+                           std::size_t core_index)
     : cb(std::make_unique<ARM_Dynarmic_Callbacks>(*this)), core_index{core_index},
       exclusive_monitor{std::dynamic_pointer_cast<DynarmicExclusiveMonitor>(exclusive_monitor)} {
     ThreadContext ctx;
@@ -161,12 +182,12 @@ ARM_Dynarmic::ARM_Dynarmic(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
 
 ARM_Dynarmic::~ARM_Dynarmic() = default;
 
-void ARM_Dynarmic::MapBackingMemory(u64 address, size_t size, u8* memory,
+void ARM_Dynarmic::MapBackingMemory(u64 address, std::size_t size, u8* memory,
                                     Kernel::VMAPermission perms) {
     inner_unicorn.MapBackingMemory(address, size, memory, perms);
 }
 
-void ARM_Dynarmic::UnmapMemory(u64 address, size_t size) {
+void ARM_Dynarmic::UnmapMemory(u64 address, std::size_t size) {
     inner_unicorn.UnmapMemory(address, size);
 }
 
@@ -186,29 +207,20 @@ void ARM_Dynarmic::SetReg(int index, u64 value) {
     jit->SetRegister(index, value);
 }
 
-u128 ARM_Dynarmic::GetExtReg(int index) const {
+u128 ARM_Dynarmic::GetVectorReg(int index) const {
     return jit->GetVector(index);
 }
 
-void ARM_Dynarmic::SetExtReg(int index, u128 value) {
+void ARM_Dynarmic::SetVectorReg(int index, u128 value) {
     jit->SetVector(index, value);
 }
 
-u32 ARM_Dynarmic::GetVFPReg(int /*index*/) const {
-    UNIMPLEMENTED();
-    return {};
-}
-
-void ARM_Dynarmic::SetVFPReg(int /*index*/, u32 /*value*/) {
-    UNIMPLEMENTED();
-}
-
-u32 ARM_Dynarmic::GetCPSR() const {
+u32 ARM_Dynarmic::GetPSTATE() const {
     return jit->GetPstate();
 }
 
-void ARM_Dynarmic::SetCPSR(u32 cpsr) {
-    jit->SetPstate(cpsr);
+void ARM_Dynarmic::SetPSTATE(u32 pstate) {
+    jit->SetPstate(pstate);
 }
 
 u64 ARM_Dynarmic::GetTlsAddress() const {
@@ -231,18 +243,18 @@ void ARM_Dynarmic::SaveContext(ThreadContext& ctx) {
     ctx.cpu_registers = jit->GetRegisters();
     ctx.sp = jit->GetSP();
     ctx.pc = jit->GetPC();
-    ctx.cpsr = jit->GetPstate();
-    ctx.fpu_registers = jit->GetVectors();
-    ctx.fpscr = jit->GetFpcr();
+    ctx.pstate = jit->GetPstate();
+    ctx.vector_registers = jit->GetVectors();
+    ctx.fpcr = jit->GetFpcr();
 }
 
 void ARM_Dynarmic::LoadContext(const ThreadContext& ctx) {
     jit->SetRegisters(ctx.cpu_registers);
     jit->SetSP(ctx.sp);
     jit->SetPC(ctx.pc);
-    jit->SetPstate(static_cast<u32>(ctx.cpsr));
-    jit->SetVectors(ctx.fpu_registers);
-    jit->SetFpcr(static_cast<u32>(ctx.fpscr));
+    jit->SetPstate(static_cast<u32>(ctx.pstate));
+    jit->SetVectors(ctx.vector_registers);
+    jit->SetFpcr(static_cast<u32>(ctx.fpcr));
 }
 
 void ARM_Dynarmic::PrepareReschedule() {
@@ -262,10 +274,10 @@ void ARM_Dynarmic::PageTableChanged() {
     current_page_table = Memory::GetCurrentPageTable();
 }
 
-DynarmicExclusiveMonitor::DynarmicExclusiveMonitor(size_t core_count) : monitor(core_count) {}
+DynarmicExclusiveMonitor::DynarmicExclusiveMonitor(std::size_t core_count) : monitor(core_count) {}
 DynarmicExclusiveMonitor::~DynarmicExclusiveMonitor() = default;
 
-void DynarmicExclusiveMonitor::SetExclusive(size_t core_index, VAddr addr) {
+void DynarmicExclusiveMonitor::SetExclusive(std::size_t core_index, VAddr addr) {
     // Size doesn't actually matter.
     monitor.Mark(core_index, addr, 16);
 }
@@ -274,29 +286,31 @@ void DynarmicExclusiveMonitor::ClearExclusive() {
     monitor.Clear();
 }
 
-bool DynarmicExclusiveMonitor::ExclusiveWrite8(size_t core_index, VAddr vaddr, u8 value) {
+bool DynarmicExclusiveMonitor::ExclusiveWrite8(std::size_t core_index, VAddr vaddr, u8 value) {
     return monitor.DoExclusiveOperation(core_index, vaddr, 1,
                                         [&] { Memory::Write8(vaddr, value); });
 }
 
-bool DynarmicExclusiveMonitor::ExclusiveWrite16(size_t core_index, VAddr vaddr, u16 value) {
+bool DynarmicExclusiveMonitor::ExclusiveWrite16(std::size_t core_index, VAddr vaddr, u16 value) {
     return monitor.DoExclusiveOperation(core_index, vaddr, 2,
                                         [&] { Memory::Write16(vaddr, value); });
 }
 
-bool DynarmicExclusiveMonitor::ExclusiveWrite32(size_t core_index, VAddr vaddr, u32 value) {
+bool DynarmicExclusiveMonitor::ExclusiveWrite32(std::size_t core_index, VAddr vaddr, u32 value) {
     return monitor.DoExclusiveOperation(core_index, vaddr, 4,
                                         [&] { Memory::Write32(vaddr, value); });
 }
 
-bool DynarmicExclusiveMonitor::ExclusiveWrite64(size_t core_index, VAddr vaddr, u64 value) {
+bool DynarmicExclusiveMonitor::ExclusiveWrite64(std::size_t core_index, VAddr vaddr, u64 value) {
     return monitor.DoExclusiveOperation(core_index, vaddr, 8,
                                         [&] { Memory::Write64(vaddr, value); });
 }
 
-bool DynarmicExclusiveMonitor::ExclusiveWrite128(size_t core_index, VAddr vaddr, u128 value) {
+bool DynarmicExclusiveMonitor::ExclusiveWrite128(std::size_t core_index, VAddr vaddr, u128 value) {
     return monitor.DoExclusiveOperation(core_index, vaddr, 16, [&] {
-        Memory::Write64(vaddr, value[0]);
-        Memory::Write64(vaddr, value[1]);
+        Memory::Write64(vaddr + 0, value[0]);
+        Memory::Write64(vaddr + 8, value[1]);
     });
 }
+
+} // namespace Core

src/core/arm/dynarmic/arm_dynarmic.h

@@ -12,29 +12,33 @@
 #include "core/arm/exclusive_monitor.h"
 #include "core/arm/unicorn/arm_unicorn.h"
 
+namespace Memory {
+struct PageTable;
+}
+
+namespace Core {
+
 class ARM_Dynarmic_Callbacks;
 class DynarmicExclusiveMonitor;
 
 class ARM_Dynarmic final : public ARM_Interface {
 public:
-    ARM_Dynarmic(std::shared_ptr<ExclusiveMonitor> exclusive_monitor, size_t core_index);
+    ARM_Dynarmic(std::shared_ptr<ExclusiveMonitor> exclusive_monitor, std::size_t core_index);
     ~ARM_Dynarmic();
 
-    void MapBackingMemory(VAddr address, size_t size, u8* memory,
+    void MapBackingMemory(VAddr address, std::size_t size, u8* memory,
                           Kernel::VMAPermission perms) override;
-    void UnmapMemory(u64 address, size_t size) override;
+    void UnmapMemory(u64 address, std::size_t size) override;
     void SetPC(u64 pc) override;
     u64 GetPC() const override;
     u64 GetReg(int index) const override;
     void SetReg(int index, u64 value) override;
-    u128 GetExtReg(int index) const override;
-    void SetExtReg(int index, u128 value) override;
-    u32 GetVFPReg(int index) const override;
-    void SetVFPReg(int index, u32 value) override;
-    u32 GetCPSR() const override;
+    u128 GetVectorReg(int index) const override;
+    void SetVectorReg(int index, u128 value) override;
+    u32 GetPSTATE() const override;
+    void SetPSTATE(u32 pstate) override;
     void Run() override;
     void Step() override;
-    void SetCPSR(u32 cpsr) override;
     VAddr GetTlsAddress() const override;
     void SetTlsAddress(VAddr address) override;
     void SetTPIDR_EL0(u64 value) override;
@@ -57,7 +61,7 @@ private:
     std::unique_ptr<Dynarmic::A64::Jit> jit;
     ARM_Unicorn inner_unicorn;
 
-    size_t core_index;
+    std::size_t core_index;
     std::shared_ptr<DynarmicExclusiveMonitor> exclusive_monitor;
 
     Memory::PageTable* current_page_table = nullptr;
@@ -65,19 +69,21 @@ private:
 
 class DynarmicExclusiveMonitor final : public ExclusiveMonitor {
 public:
-    explicit DynarmicExclusiveMonitor(size_t core_count);
+    explicit DynarmicExclusiveMonitor(std::size_t core_count);
     ~DynarmicExclusiveMonitor();
 
-    void SetExclusive(size_t core_index, VAddr addr) override;
+    void SetExclusive(std::size_t core_index, VAddr addr) override;
     void ClearExclusive() override;
 
-    bool ExclusiveWrite8(size_t core_index, VAddr vaddr, u8 value) override;
-    bool ExclusiveWrite16(size_t core_index, VAddr vaddr, u16 value) override;
-    bool ExclusiveWrite32(size_t core_index, VAddr vaddr, u32 value) override;
-    bool ExclusiveWrite64(size_t core_index, VAddr vaddr, u64 value) override;
-    bool ExclusiveWrite128(size_t core_index, VAddr vaddr, u128 value) override;
+    bool ExclusiveWrite8(std::size_t core_index, VAddr vaddr, u8 value) override;
+    bool ExclusiveWrite16(std::size_t core_index, VAddr vaddr, u16 value) override;
+    bool ExclusiveWrite32(std::size_t core_index, VAddr vaddr, u32 value) override;
+    bool ExclusiveWrite64(std::size_t core_index, VAddr vaddr, u64 value) override;
+    bool ExclusiveWrite128(std::size_t core_index, VAddr vaddr, u128 value) override;
 
 private:
     friend class ARM_Dynarmic;
     Dynarmic::A64::ExclusiveMonitor monitor;
 };
+
+} // namespace Core

src/core/arm/exclusive_monitor.cpp

@@ -4,4 +4,8 @@
 
 #include "core/arm/exclusive_monitor.h"
 
+namespace Core {
+
 ExclusiveMonitor::~ExclusiveMonitor() = default;
+
+} // namespace Core

src/core/arm/exclusive_monitor.h

@@ -6,16 +6,20 @@
 
 #include "common/common_types.h"
 
+namespace Core {
+
 class ExclusiveMonitor {
 public:
     virtual ~ExclusiveMonitor();
 
-    virtual void SetExclusive(size_t core_index, VAddr addr) = 0;
+    virtual void SetExclusive(std::size_t core_index, VAddr addr) = 0;
     virtual void ClearExclusive() = 0;
 
-    virtual bool ExclusiveWrite8(size_t core_index, VAddr vaddr, u8 value) = 0;
-    virtual bool ExclusiveWrite16(size_t core_index, VAddr vaddr, u16 value) = 0;
-    virtual bool ExclusiveWrite32(size_t core_index, VAddr vaddr, u32 value) = 0;
-    virtual bool ExclusiveWrite64(size_t core_index, VAddr vaddr, u64 value) = 0;
-    virtual bool ExclusiveWrite128(size_t core_index, VAddr vaddr, u128 value) = 0;
+    virtual bool ExclusiveWrite8(std::size_t core_index, VAddr vaddr, u8 value) = 0;
+    virtual bool ExclusiveWrite16(std::size_t core_index, VAddr vaddr, u16 value) = 0;
+    virtual bool ExclusiveWrite32(std::size_t core_index, VAddr vaddr, u32 value) = 0;
+    virtual bool ExclusiveWrite64(std::size_t core_index, VAddr vaddr, u64 value) = 0;
+    virtual bool ExclusiveWrite128(std::size_t core_index, VAddr vaddr, u128 value) = 0;
 };
+
+} // namespace Core

src/core/arm/unicorn/arm_unicorn.cpp

@@ -11,6 +11,8 @@
 #include "core/core_timing.h"
 #include "core/hle/kernel/svc.h"
 
+namespace Core {
+
 // Load Unicorn DLL once on Windows using RAII
 #ifdef _MSC_VER
 #include <unicorn_dynload.h>
@@ -88,12 +90,12 @@ ARM_Unicorn::~ARM_Unicorn() {
     CHECKED(uc_close(uc));
 }
 
-void ARM_Unicorn::MapBackingMemory(VAddr address, size_t size, u8* memory,
+void ARM_Unicorn::MapBackingMemory(VAddr address, std::size_t size, u8* memory,
                                    Kernel::VMAPermission perms) {
     CHECKED(uc_mem_map_ptr(uc, address, size, static_cast<u32>(perms), memory));
 }
 
-void ARM_Unicorn::UnmapMemory(VAddr address, size_t size) {
+void ARM_Unicorn::UnmapMemory(VAddr address, std::size_t size) {
     CHECKED(uc_mem_unmap(uc, address, size));
 }
 
@@ -129,33 +131,24 @@ void ARM_Unicorn::SetReg(int regn, u64 val) {
     CHECKED(uc_reg_write(uc, treg, &val));
 }
 
-u128 ARM_Unicorn::GetExtReg(int /*index*/) const {
+u128 ARM_Unicorn::GetVectorReg(int /*index*/) const {
     UNIMPLEMENTED();
     static constexpr u128 res{};
     return res;
 }
 
-void ARM_Unicorn::SetExtReg(int /*index*/, u128 /*value*/) {
+void ARM_Unicorn::SetVectorReg(int /*index*/, u128 /*value*/) {
     UNIMPLEMENTED();
 }
 
-u32 ARM_Unicorn::GetVFPReg(int /*index*/) const {
-    UNIMPLEMENTED();
-    return {};
-}
-
-void ARM_Unicorn::SetVFPReg(int /*index*/, u32 /*value*/) {
-    UNIMPLEMENTED();
-}
-
-u32 ARM_Unicorn::GetCPSR() const {
+u32 ARM_Unicorn::GetPSTATE() const {
     u64 nzcv{};
     CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &nzcv));
     return static_cast<u32>(nzcv);
 }
 
-void ARM_Unicorn::SetCPSR(u32 cpsr) {
-    u64 nzcv = cpsr;
+void ARM_Unicorn::SetPSTATE(u32 pstate) {
+    u64 nzcv = pstate;
     CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &nzcv));
 }
 
@@ -191,10 +184,10 @@ void ARM_Unicorn::Step() {
     ExecuteInstructions(1);
 }
 
-MICROPROFILE_DEFINE(ARM_Jit, "ARM JIT", "ARM JIT", MP_RGB(255, 64, 64));
+MICROPROFILE_DEFINE(ARM_Jit_Unicorn, "ARM JIT", "Unicorn", MP_RGB(255, 64, 64));
 
 void ARM_Unicorn::ExecuteInstructions(int num_instructions) {
-    MICROPROFILE_SCOPE(ARM_Jit);
+    MICROPROFILE_SCOPE(ARM_Jit_Unicorn);
     CHECKED(uc_emu_start(uc, GetPC(), 1ULL << 63, 0, num_instructions));
     CoreTiming::AddTicks(num_instructions);
     if (GDBStub::IsServerEnabled()) {
@@ -211,13 +204,13 @@ void ARM_Unicorn::ExecuteInstructions(int num_instructions) {
     }
 }
 
-void ARM_Unicorn::SaveContext(ARM_Interface::ThreadContext& ctx) {
+void ARM_Unicorn::SaveContext(ThreadContext& ctx) {
     int uregs[32];
     void* tregs[32];
 
     CHECKED(uc_reg_read(uc, UC_ARM64_REG_SP, &ctx.sp));
     CHECKED(uc_reg_read(uc, UC_ARM64_REG_PC, &ctx.pc));
-    CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &ctx.cpsr));
+    CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &ctx.pstate));
 
     for (auto i = 0; i < 29; ++i) {
         uregs[i] = UC_ARM64_REG_X0 + i;
@@ -232,19 +225,19 @@ void ARM_Unicorn::SaveContext(ARM_Interface::ThreadContext& ctx) {
 
     for (int i = 0; i < 32; ++i) {
         uregs[i] = UC_ARM64_REG_Q0 + i;
-        tregs[i] = &ctx.fpu_registers[i];
+        tregs[i] = &ctx.vector_registers[i];
     }
 
     CHECKED(uc_reg_read_batch(uc, uregs, tregs, 32));
 }
 
-void ARM_Unicorn::LoadContext(const ARM_Interface::ThreadContext& ctx) {
+void ARM_Unicorn::LoadContext(const ThreadContext& ctx) {
     int uregs[32];
     void* tregs[32];
 
     CHECKED(uc_reg_write(uc, UC_ARM64_REG_SP, &ctx.sp));
     CHECKED(uc_reg_write(uc, UC_ARM64_REG_PC, &ctx.pc));
-    CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &ctx.cpsr));
+    CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &ctx.pstate));
 
     for (int i = 0; i < 29; ++i) {
         uregs[i] = UC_ARM64_REG_X0 + i;
@@ -259,7 +252,7 @@ void ARM_Unicorn::LoadContext(const ARM_Interface::ThreadContext& ctx) {
 
     for (auto i = 0; i < 32; ++i) {
         uregs[i] = UC_ARM64_REG_Q0 + i;
-        tregs[i] = (void*)&ctx.fpu_registers[i];
+        tregs[i] = (void*)&ctx.vector_registers[i];
     }
 
     CHECKED(uc_reg_write_batch(uc, uregs, tregs, 32));
@@ -277,3 +270,5 @@ void ARM_Unicorn::RecordBreak(GDBStub::BreakpointAddress bkpt) {
     last_bkpt = bkpt;
     last_bkpt_hit = true;
 }
+
+} // namespace Core

src/core/arm/unicorn/arm_unicorn.h

@@ -9,23 +9,23 @@
 #include "core/arm/arm_interface.h"
 #include "core/gdbstub/gdbstub.h"
 
+namespace Core {
+
 class ARM_Unicorn final : public ARM_Interface {
 public:
     ARM_Unicorn();
     ~ARM_Unicorn();
-    void MapBackingMemory(VAddr address, size_t size, u8* memory,
+    void MapBackingMemory(VAddr address, std::size_t size, u8* memory,
                           Kernel::VMAPermission perms) override;
-    void UnmapMemory(VAddr address, size_t size) override;
+    void UnmapMemory(VAddr address, std::size_t size) override;
     void SetPC(u64 pc) override;
     u64 GetPC() const override;
     u64 GetReg(int index) const override;
     void SetReg(int index, u64 value) override;
-    u128 GetExtReg(int index) const override;
-    void SetExtReg(int index, u128 value) override;
-    u32 GetVFPReg(int index) const override;
-    void SetVFPReg(int index, u32 value) override;
-    u32 GetCPSR() const override;
-    void SetCPSR(u32 cpsr) override;
+    u128 GetVectorReg(int index) const override;
+    void SetVectorReg(int index, u128 value) override;
+    u32 GetPSTATE() const override;
+    void SetPSTATE(u32 pstate) override;
     VAddr GetTlsAddress() const override;
     void SetTlsAddress(VAddr address) override;
     void SetTPIDR_EL0(u64 value) override;
@@ -46,3 +46,5 @@ private:
     GDBStub::BreakpointAddress last_bkpt{};
     bool last_bkpt_hit;
 };
+
+} // namespace Core

src/core/core.cpp

@@ -2,24 +2,35 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <array>
+#include <map>
 #include <memory>
+#include <thread>
 #include <utility>
+
 #include "common/logging/log.h"
 #include "common/string_util.h"
+#include "core/arm/exclusive_monitor.h"
 #include "core/core.h"
+#include "core/core_cpu.h"
 #include "core/core_timing.h"
+#include "core/file_sys/mode.h"
+#include "core/file_sys/vfs_concat.h"
+#include "core/file_sys/vfs_real.h"
 #include "core/gdbstub/gdbstub.h"
 #include "core/hle/kernel/client_port.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
+#include "core/hle/kernel/scheduler.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/service/service.h"
-#include "core/hle/service/sm/controller.h"
 #include "core/hle/service/sm/sm.h"
 #include "core/loader/loader.h"
+#include "core/perf_stats.h"
 #include "core/settings.h"
-#include "file_sys/vfs_concat.h"
-#include "file_sys/vfs_real.h"
+#include "core/telemetry_session.h"
+#include "video_core/debug_utils/debug_utils.h"
+#include "video_core/gpu.h"
 #include "video_core/renderer_base.h"
 #include "video_core/video_core.h"
 
@@ -27,71 +38,9 @@ namespace Core {
 
 /*static*/ System System::s_instance;
 
-System::System() = default;
-
-System::~System() = default;
-
-/// Runs a CPU core while the system is powered on
-static void RunCpuCore(std::shared_ptr<Cpu> cpu_state) {
-    while (Core::System::GetInstance().IsPoweredOn()) {
-        cpu_state->RunLoop(true);
-    }
-}
-
-Cpu& System::CurrentCpuCore() {
-    // If multicore is enabled, use host thread to figure out the current CPU core
-    if (Settings::values.use_multi_core) {
-        const auto& search = thread_to_cpu.find(std::this_thread::get_id());
-        ASSERT(search != thread_to_cpu.end());
-        ASSERT(search->second);
-        return *search->second;
-    }
-
-    // Otherwise, use single-threaded mode active_core variable
-    return *cpu_cores[active_core];
-}
-
-System::ResultStatus System::RunLoop(bool tight_loop) {
-    status = ResultStatus::Success;
-
-    // Update thread_to_cpu in case Core 0 is run from a different host thread
-    thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
-
-    if (GDBStub::IsServerEnabled()) {
-        GDBStub::HandlePacket();
-
-        // If the loop is halted and we want to step, use a tiny (1) number of instructions to
-        // execute. Otherwise, get out of the loop function.
-        if (GDBStub::GetCpuHaltFlag()) {
-            if (GDBStub::GetCpuStepFlag()) {
-                tight_loop = false;
-            } else {
-                return ResultStatus::Success;
-            }
-        }
-    }
-
-    for (active_core = 0; active_core < NUM_CPU_CORES; ++active_core) {
-        cpu_cores[active_core]->RunLoop(tight_loop);
-        if (Settings::values.use_multi_core) {
-            // Cores 1-3 are run on other threads in this mode
-            break;
-        }
-    }
-
-    if (GDBStub::IsServerEnabled()) {
-        GDBStub::SetCpuStepFlag(false);
-    }
-
-    return status;
-}
-
-System::ResultStatus System::SingleStep() {
-    return RunLoop(false);
-}
-
-static FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
-                                                const std::string& path) {
+namespace {
+FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
+                                         const std::string& path) {
     // To account for split 00+01+etc files.
     std::string dir_name;
     std::string filename;
@@ -121,168 +70,405 @@ static FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem
     return vfs->OpenFile(path, FileSys::Mode::Read);
 }
 
-System::ResultStatus System::Load(Frontend::EmuWindow& emu_window, const std::string& filepath) {
-    app_loader = Loader::GetLoader(GetGameFileFromPath(virtual_filesystem, filepath));
-
-    if (!app_loader) {
-        LOG_CRITICAL(Core, "Failed to obtain loader for {}!", filepath);
-        return ResultStatus::ErrorGetLoader;
+/// Runs a CPU core while the system is powered on
+void RunCpuCore(std::shared_ptr<Cpu> cpu_state) {
+    while (Core::System::GetInstance().IsPoweredOn()) {
+        cpu_state->RunLoop(true);
     }
-    std::pair<boost::optional<u32>, Loader::ResultStatus> system_mode =
-        app_loader->LoadKernelSystemMode();
+}
+} // Anonymous namespace
 
-    if (system_mode.second != Loader::ResultStatus::Success) {
-        LOG_CRITICAL(Core, "Failed to determine system mode (Error {})!",
-                     static_cast<int>(system_mode.second));
+struct System::Impl {
+    Cpu& CurrentCpuCore() {
+        if (Settings::values.use_multi_core) {
+            const auto& search = thread_to_cpu.find(std::this_thread::get_id());
+            ASSERT(search != thread_to_cpu.end());
+            ASSERT(search->second);
+            return *search->second;
+        }
+
+        // Otherwise, use single-threaded mode active_core variable
+        return *cpu_cores[active_core];
+    }
+
+    ResultStatus RunLoop(bool tight_loop) {
+        status = ResultStatus::Success;
+
+        // Update thread_to_cpu in case Core 0 is run from a different host thread
+        thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
+
+        if (GDBStub::IsServerEnabled()) {
+            GDBStub::HandlePacket();
+
+            // If the loop is halted and we want to step, use a tiny (1) number of instructions to
+            // execute. Otherwise, get out of the loop function.
+            if (GDBStub::GetCpuHaltFlag()) {
+                if (GDBStub::GetCpuStepFlag()) {
+                    tight_loop = false;
+                } else {
+                    return ResultStatus::Success;
+                }
+            }
+        }
+
+        for (active_core = 0; active_core < NUM_CPU_CORES; ++active_core) {
+            cpu_cores[active_core]->RunLoop(tight_loop);
+            if (Settings::values.use_multi_core) {
+                // Cores 1-3 are run on other threads in this mode
+                break;
+            }
+        }
+
+        if (GDBStub::IsServerEnabled()) {
+            GDBStub::SetCpuStepFlag(false);
+        }
+
+        return status;
+    }
+
+    ResultStatus Init(Frontend::EmuWindow& emu_window) {
+        LOG_DEBUG(HW_Memory, "initialized OK");
+
+        CoreTiming::Init();
+        kernel.Initialize();
+
+        // Create a default fs if one doesn't already exist.
+        if (virtual_filesystem == nullptr)
+            virtual_filesystem = std::make_shared<FileSys::RealVfsFilesystem>();
+
+        kernel.MakeCurrentProcess(Kernel::Process::Create(kernel, "main"));
+
+        cpu_barrier = std::make_shared<CpuBarrier>();
+        cpu_exclusive_monitor = Cpu::MakeExclusiveMonitor(cpu_cores.size());
+        for (std::size_t index = 0; index < cpu_cores.size(); ++index) {
+            cpu_cores[index] = std::make_shared<Cpu>(cpu_exclusive_monitor, cpu_barrier, index);
+        }
+
+        telemetry_session = std::make_unique<Core::TelemetrySession>();
+        service_manager = std::make_shared<Service::SM::ServiceManager>();
+
+        Service::Init(service_manager, virtual_filesystem);
+        GDBStub::Init();
+
+        renderer = VideoCore::CreateRenderer(emu_window);
+        if (!renderer->Init()) {
+            return ResultStatus::ErrorVideoCore;
+        }
+
+        gpu_core = std::make_unique<Tegra::GPU>(renderer->Rasterizer());
+
+        // Create threads for CPU cores 1-3, and build thread_to_cpu map
+        // CPU core 0 is run on the main thread
+        thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
+        if (Settings::values.use_multi_core) {
+            for (std::size_t index = 0; index < cpu_core_threads.size(); ++index) {
+                cpu_core_threads[index] =
+                    std::make_unique<std::thread>(RunCpuCore, cpu_cores[index + 1]);
+                thread_to_cpu[cpu_core_threads[index]->get_id()] = cpu_cores[index + 1];
+            }
+        }
+
+        LOG_DEBUG(Core, "Initialized OK");
+
+        // Reset counters and set time origin to current frame
+        GetAndResetPerfStats();
+        perf_stats.BeginSystemFrame();
+
+        return ResultStatus::Success;
+    }
+
+    ResultStatus Load(Frontend::EmuWindow& emu_window, const std::string& filepath) {
+        app_loader = Loader::GetLoader(GetGameFileFromPath(virtual_filesystem, filepath));
+
+        if (!app_loader) {
+            LOG_CRITICAL(Core, "Failed to obtain loader for {}!", filepath);
+            return ResultStatus::ErrorGetLoader;
+        }
+        std::pair<boost::optional<u32>, Loader::ResultStatus> system_mode =
+            app_loader->LoadKernelSystemMode();
+
+        if (system_mode.second != Loader::ResultStatus::Success) {
+            LOG_CRITICAL(Core, "Failed to determine system mode (Error {})!",
+                         static_cast<int>(system_mode.second));
 
-        if (system_mode.second != Loader::ResultStatus::Success)
             return ResultStatus::ErrorSystemMode;
-    }
+        }
 
-    ResultStatus init_result{Init(emu_window)};
-    if (init_result != ResultStatus::Success) {
-        LOG_CRITICAL(Core, "Failed to initialize system (Error {})!",
-                     static_cast<int>(init_result));
-        System::Shutdown();
-        return init_result;
-    }
-
-    const Loader::ResultStatus load_result{app_loader->Load(current_process)};
-    if (Loader::ResultStatus::Success != load_result) {
-        LOG_CRITICAL(Core, "Failed to load ROM (Error {})!", static_cast<int>(load_result));
-        System::Shutdown();
+        ResultStatus init_result{Init(emu_window)};
+        if (init_result != ResultStatus::Success) {
+            LOG_CRITICAL(Core, "Failed to initialize system (Error {})!",
+                         static_cast<int>(init_result));
+            Shutdown();
+            return init_result;
+        }
 
+        const Loader::ResultStatus load_result{app_loader->Load(kernel.CurrentProcess())};
         if (load_result != Loader::ResultStatus::Success) {
+            LOG_CRITICAL(Core, "Failed to load ROM (Error {})!", static_cast<int>(load_result));
+            Shutdown();
+
             return static_cast<ResultStatus>(static_cast<u32>(ResultStatus::ErrorLoader) +
                                              static_cast<u32>(load_result));
         }
+        status = ResultStatus::Success;
+        return status;
     }
-    status = ResultStatus::Success;
-    return status;
+
+    void Shutdown() {
+        // Log last frame performance stats
+        auto perf_results = GetAndResetPerfStats();
+        Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_EmulationSpeed",
+                             perf_results.emulation_speed * 100.0);
+        Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_Framerate",
+                             perf_results.game_fps);
+        Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_Frametime",
+                             perf_results.frametime * 1000.0);
+
+        // Shutdown emulation session
+        renderer.reset();
+        GDBStub::Shutdown();
+        Service::Shutdown();
+        service_manager.reset();
+        telemetry_session.reset();
+        gpu_core.reset();
+
+        // Close all CPU/threading state
+        cpu_barrier->NotifyEnd();
+        if (Settings::values.use_multi_core) {
+            for (auto& thread : cpu_core_threads) {
+                thread->join();
+                thread.reset();
+            }
+        }
+        thread_to_cpu.clear();
+        for (auto& cpu_core : cpu_cores) {
+            cpu_core.reset();
+        }
+        cpu_barrier.reset();
+
+        // Shutdown kernel and core timing
+        kernel.Shutdown();
+        CoreTiming::Shutdown();
+
+        // Close app loader
+        app_loader.reset();
+
+        LOG_DEBUG(Core, "Shutdown OK");
+    }
+
+    Loader::ResultStatus GetGameName(std::string& out) const {
+        if (app_loader == nullptr)
+            return Loader::ResultStatus::ErrorNotInitialized;
+        return app_loader->ReadTitle(out);
+    }
+
+    void SetStatus(ResultStatus new_status, const char* details = nullptr) {
+        status = new_status;
+        if (details) {
+            status_details = details;
+        }
+    }
+
+    PerfStatsResults GetAndResetPerfStats() {
+        return perf_stats.GetAndResetStats(CoreTiming::GetGlobalTimeUs());
+    }
+
+    Kernel::KernelCore kernel;
+    /// RealVfsFilesystem instance
+    FileSys::VirtualFilesystem virtual_filesystem;
+    /// AppLoader used to load the current executing application
+    std::unique_ptr<Loader::AppLoader> app_loader;
+    std::unique_ptr<VideoCore::RendererBase> renderer;
+    std::unique_ptr<Tegra::GPU> gpu_core;
+    std::shared_ptr<Tegra::DebugContext> debug_context;
+    std::shared_ptr<ExclusiveMonitor> cpu_exclusive_monitor;
+    std::shared_ptr<CpuBarrier> cpu_barrier;
+    std::array<std::shared_ptr<Cpu>, NUM_CPU_CORES> cpu_cores;
+    std::array<std::unique_ptr<std::thread>, NUM_CPU_CORES - 1> cpu_core_threads;
+    std::size_t active_core{}; ///< Active core, only used in single thread mode
+
+    /// Service manager
+    std::shared_ptr<Service::SM::ServiceManager> service_manager;
+
+    /// Telemetry session for this emulation session
+    std::unique_ptr<Core::TelemetrySession> telemetry_session;
+
+    ResultStatus status = ResultStatus::Success;
+    std::string status_details = "";
+
+    /// Map of guest threads to CPU cores
+    std::map<std::thread::id, std::shared_ptr<Cpu>> thread_to_cpu;
+
+    Core::PerfStats perf_stats;
+    Core::FrameLimiter frame_limiter;
+};
+
+System::System() : impl{std::make_unique<Impl>()} {}
+System::~System() = default;
+
+Cpu& System::CurrentCpuCore() {
+    return impl->CurrentCpuCore();
+}
+
+System::ResultStatus System::RunLoop(bool tight_loop) {
+    return impl->RunLoop(tight_loop);
+}
+
+System::ResultStatus System::SingleStep() {
+    return RunLoop(false);
+}
+
+void System::InvalidateCpuInstructionCaches() {
+    for (auto& cpu : impl->cpu_cores) {
+        cpu->ArmInterface().ClearInstructionCache();
+    }
+}
+
+System::ResultStatus System::Load(Frontend::EmuWindow& emu_window, const std::string& filepath) {
+    return impl->Load(emu_window, filepath);
+}
+
+bool System::IsPoweredOn() const {
+    return impl->cpu_barrier && impl->cpu_barrier->IsAlive();
 }
 
 void System::PrepareReschedule() {
     CurrentCpuCore().PrepareReschedule();
 }
 
-PerfStats::Results System::GetAndResetPerfStats() {
-    return perf_stats.GetAndResetStats(CoreTiming::GetGlobalTimeUs());
+PerfStatsResults System::GetAndResetPerfStats() {
+    return impl->GetAndResetPerfStats();
 }
 
-const std::shared_ptr<Kernel::Scheduler>& System::Scheduler(size_t core_index) {
-    ASSERT(core_index < NUM_CPU_CORES);
-    return cpu_cores[core_index]->Scheduler();
+Core::TelemetrySession& System::TelemetrySession() const {
+    return *impl->telemetry_session;
 }
 
-ARM_Interface& System::ArmInterface(size_t core_index) {
-    ASSERT(core_index < NUM_CPU_CORES);
-    return cpu_cores[core_index]->ArmInterface();
+ARM_Interface& System::CurrentArmInterface() {
+    return CurrentCpuCore().ArmInterface();
 }
 
-Cpu& System::CpuCore(size_t core_index) {
+std::size_t System::CurrentCoreIndex() {
+    return CurrentCpuCore().CoreIndex();
+}
+
+Kernel::Scheduler& System::CurrentScheduler() {
+    return *CurrentCpuCore().Scheduler();
+}
+
+const std::shared_ptr<Kernel::Scheduler>& System::Scheduler(std::size_t core_index) {
     ASSERT(core_index < NUM_CPU_CORES);
-    return *cpu_cores[core_index];
+    return impl->cpu_cores[core_index]->Scheduler();
+}
+
+Kernel::SharedPtr<Kernel::Process>& System::CurrentProcess() {
+    return impl->kernel.CurrentProcess();
+}
+
+const Kernel::SharedPtr<Kernel::Process>& System::CurrentProcess() const {
+    return impl->kernel.CurrentProcess();
+}
+
+ARM_Interface& System::ArmInterface(std::size_t core_index) {
+    ASSERT(core_index < NUM_CPU_CORES);
+    return impl->cpu_cores[core_index]->ArmInterface();
+}
+
+Cpu& System::CpuCore(std::size_t core_index) {
+    ASSERT(core_index < NUM_CPU_CORES);
+    return *impl->cpu_cores[core_index];
+}
+
+ExclusiveMonitor& System::Monitor() {
+    return *impl->cpu_exclusive_monitor;
+}
+
+Tegra::GPU& System::GPU() {
+    return *impl->gpu_core;
+}
+
+const Tegra::GPU& System::GPU() const {
+    return *impl->gpu_core;
+}
+
+VideoCore::RendererBase& System::Renderer() {
+    return *impl->renderer;
+}
+
+const VideoCore::RendererBase& System::Renderer() const {
+    return *impl->renderer;
+}
+
+Kernel::KernelCore& System::Kernel() {
+    return impl->kernel;
+}
+
+const Kernel::KernelCore& System::Kernel() const {
+    return impl->kernel;
+}
+
+Core::PerfStats& System::GetPerfStats() {
+    return impl->perf_stats;
+}
+
+const Core::PerfStats& System::GetPerfStats() const {
+    return impl->perf_stats;
+}
+
+Core::FrameLimiter& System::FrameLimiter() {
+    return impl->frame_limiter;
+}
+
+const Core::FrameLimiter& System::FrameLimiter() const {
+    return impl->frame_limiter;
+}
+
+Loader::ResultStatus System::GetGameName(std::string& out) const {
+    return impl->GetGameName(out);
+}
+
+void System::SetStatus(ResultStatus new_status, const char* details) {
+    impl->SetStatus(new_status, details);
+}
+
+const std::string& System::GetStatusDetails() const {
+    return impl->status_details;
+}
+
+Loader::AppLoader& System::GetAppLoader() const {
+    return *impl->app_loader;
+}
+
+void System::SetGPUDebugContext(std::shared_ptr<Tegra::DebugContext> context) {
+    impl->debug_context = std::move(context);
+}
+
+Tegra::DebugContext* System::GetGPUDebugContext() const {
+    return impl->debug_context.get();
+}
+
+void System::SetFilesystem(std::shared_ptr<FileSys::VfsFilesystem> vfs) {
+    impl->virtual_filesystem = std::move(vfs);
+}
+
+std::shared_ptr<FileSys::VfsFilesystem> System::GetFilesystem() const {
+    return impl->virtual_filesystem;
 }
 
 System::ResultStatus System::Init(Frontend::EmuWindow& emu_window) {
-    LOG_DEBUG(HW_Memory, "initialized OK");
-
-    CoreTiming::Init();
-
-    // Create a default fs if one doesn't already exist.
-    if (virtual_filesystem == nullptr)
-        virtual_filesystem = std::make_shared<FileSys::RealVfsFilesystem>();
-
-    current_process = Kernel::Process::Create("main");
-
-    cpu_barrier = std::make_shared<CpuBarrier>();
-    cpu_exclusive_monitor = Cpu::MakeExclusiveMonitor(cpu_cores.size());
-    for (size_t index = 0; index < cpu_cores.size(); ++index) {
-        cpu_cores[index] = std::make_shared<Cpu>(cpu_exclusive_monitor, cpu_barrier, index);
-    }
-
-    telemetry_session = std::make_unique<Core::TelemetrySession>();
-    service_manager = std::make_shared<Service::SM::ServiceManager>();
-
-    Kernel::Init();
-    Service::Init(service_manager, virtual_filesystem);
-    GDBStub::Init();
-
-    renderer = VideoCore::CreateRenderer(emu_window);
-    if (!renderer->Init()) {
-        return ResultStatus::ErrorVideoCore;
-    }
-
-    gpu_core = std::make_unique<Tegra::GPU>(renderer->Rasterizer());
-
-    // Create threads for CPU cores 1-3, and build thread_to_cpu map
-    // CPU core 0 is run on the main thread
-    thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
-    if (Settings::values.use_multi_core) {
-        for (size_t index = 0; index < cpu_core_threads.size(); ++index) {
-            cpu_core_threads[index] =
-                std::make_unique<std::thread>(RunCpuCore, cpu_cores[index + 1]);
-            thread_to_cpu[cpu_core_threads[index]->get_id()] = cpu_cores[index + 1];
-        }
-    }
-
-    LOG_DEBUG(Core, "Initialized OK");
-
-    // Reset counters and set time origin to current frame
-    GetAndResetPerfStats();
-    perf_stats.BeginSystemFrame();
-
-    return ResultStatus::Success;
+    return impl->Init(emu_window);
 }
 
 void System::Shutdown() {
-    // Log last frame performance stats
-    auto perf_results = GetAndResetPerfStats();
-    Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_EmulationSpeed",
-                         perf_results.emulation_speed * 100.0);
-    Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_Framerate",
-                         perf_results.game_fps);
-    Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_Frametime",
-                         perf_results.frametime * 1000.0);
-
-    // Shutdown emulation session
-    renderer.reset();
-    GDBStub::Shutdown();
-    Service::Shutdown();
-    Kernel::Shutdown();
-    service_manager.reset();
-    telemetry_session.reset();
-    gpu_core.reset();
-
-    // Close all CPU/threading state
-    cpu_barrier->NotifyEnd();
-    if (Settings::values.use_multi_core) {
-        for (auto& thread : cpu_core_threads) {
-            thread->join();
-            thread.reset();
-        }
-    }
-    thread_to_cpu.clear();
-    for (auto& cpu_core : cpu_cores) {
-        cpu_core.reset();
-    }
-    cpu_barrier.reset();
-
-    // Close core timing
-    CoreTiming::Shutdown();
-
-    // Close app loader
-    app_loader.reset();
-
-    LOG_DEBUG(Core, "Shutdown OK");
+    impl->Shutdown();
 }
 
 Service::SM::ServiceManager& System::ServiceManager() {
-    return *service_manager;
+    return *impl->service_manager;
 }
 
 const Service::SM::ServiceManager& System::ServiceManager() const {
-    return *service_manager;
+    return *impl->service_manager;
 }
 
 } // namespace Core

src/core/core.h

@@ -4,40 +4,56 @@
 
 #pragma once
 
-#include <array>
+#include <cstddef>
 #include <memory>
 #include <string>
-#include <thread>
-#include "common/common_types.h"
-#include "core/arm/exclusive_monitor.h"
-#include "core/core_cpu.h"
-#include "core/hle/kernel/object.h"
-#include "core/hle/kernel/scheduler.h"
-#include "core/loader/loader.h"
-#include "core/memory.h"
-#include "core/perf_stats.h"
-#include "core/telemetry_session.h"
-#include "file_sys/vfs_real.h"
-#include "hle/service/filesystem/filesystem.h"
-#include "video_core/debug_utils/debug_utils.h"
-#include "video_core/gpu.h"
 
-class ARM_Interface;
+#include "common/common_types.h"
+#include "core/hle/kernel/object.h"
 
 namespace Core::Frontend {
 class EmuWindow;
-}
+} // namespace Core::Frontend
+
+namespace FileSys {
+class VfsFilesystem;
+} // namespace FileSys
+
+namespace Kernel {
+class KernelCore;
+class Process;
+class Scheduler;
+} // namespace Kernel
+
+namespace Loader {
+class AppLoader;
+enum class ResultStatus : u16;
+} // namespace Loader
 
 namespace Service::SM {
 class ServiceManager;
-}
+} // namespace Service::SM
+
+namespace Tegra {
+class DebugContext;
+class GPU;
+} // namespace Tegra
 
 namespace VideoCore {
 class RendererBase;
-}
+} // namespace VideoCore
 
 namespace Core {
 
+class ARM_Interface;
+class Cpu;
+class ExclusiveMonitor;
+class FrameLimiter;
+class PerfStats;
+class TelemetrySession;
+
+struct PerfStatsResults;
+
 class System {
 public:
     System(const System&) = delete;
@@ -92,11 +108,7 @@ public:
      * This function should only be used by GDB Stub to support breakpoints, memory updates and
      * step/continue commands.
      */
-    void InvalidateCpuInstructionCaches() {
-        for (auto& cpu : cpu_cores) {
-            cpu->ArmInterface().ClearInstructionCache();
-        }
-    }
+    void InvalidateCpuInstructionCaches();
 
     /// Shutdown the emulated system.
     void Shutdown();
@@ -115,114 +127,96 @@ public:
      * application).
      * @returns True if the emulated system is powered on, otherwise false.
      */
-    bool IsPoweredOn() const {
-        return cpu_barrier && cpu_barrier->IsAlive();
-    }
+    bool IsPoweredOn() const;
 
     /**
      * Returns a reference to the telemetry session for this emulation session.
      * @returns Reference to the telemetry session.
      */
-    Core::TelemetrySession& TelemetrySession() const {
-        return *telemetry_session;
-    }
+    Core::TelemetrySession& TelemetrySession() const;
 
     /// Prepare the core emulation for a reschedule
     void PrepareReschedule();
 
     /// Gets and resets core performance statistics
-    PerfStats::Results GetAndResetPerfStats();
+    PerfStatsResults GetAndResetPerfStats();
 
     /// Gets an ARM interface to the CPU core that is currently running
-    ARM_Interface& CurrentArmInterface() {
-        return CurrentCpuCore().ArmInterface();
-    }
+    ARM_Interface& CurrentArmInterface();
 
     /// Gets the index of the currently running CPU core
-    size_t CurrentCoreIndex() {
-        return CurrentCpuCore().CoreIndex();
-    }
-
-    /// Gets an ARM interface to the CPU core with the specified index
-    ARM_Interface& ArmInterface(size_t core_index);
-
-    /// Gets a CPU interface to the CPU core with the specified index
-    Cpu& CpuCore(size_t core_index);
-
-    /// Gets a mutable reference to the GPU interface
-    Tegra::GPU& GPU() {
-        return *gpu_core;
-    }
-
-    /// Gets an immutable reference to the GPU interface.
-    const Tegra::GPU& GPU() const {
-        return *gpu_core;
-    }
-
-    /// Gets a mutable reference to the renderer.
-    VideoCore::RendererBase& Renderer() {
-        return *renderer;
-    }
-
-    /// Gets an immutable reference to the renderer.
-    const VideoCore::RendererBase& Renderer() const {
-        return *renderer;
-    }
+    std::size_t CurrentCoreIndex();
 
     /// Gets the scheduler for the CPU core that is currently running
-    Kernel::Scheduler& CurrentScheduler() {
-        return *CurrentCpuCore().Scheduler();
-    }
+    Kernel::Scheduler& CurrentScheduler();
+
+    /// Gets an ARM interface to the CPU core with the specified index
+    ARM_Interface& ArmInterface(std::size_t core_index);
+
+    /// Gets a CPU interface to the CPU core with the specified index
+    Cpu& CpuCore(std::size_t core_index);
 
     /// Gets the exclusive monitor
-    ExclusiveMonitor& Monitor() {
-        return *cpu_exclusive_monitor;
-    }
+    ExclusiveMonitor& Monitor();
+
+    /// Gets a mutable reference to the GPU interface
+    Tegra::GPU& GPU();
+
+    /// Gets an immutable reference to the GPU interface.
+    const Tegra::GPU& GPU() const;
+
+    /// Gets a mutable reference to the renderer.
+    VideoCore::RendererBase& Renderer();
+
+    /// Gets an immutable reference to the renderer.
+    const VideoCore::RendererBase& Renderer() const;
 
     /// Gets the scheduler for the CPU core with the specified index
-    const std::shared_ptr<Kernel::Scheduler>& Scheduler(size_t core_index);
+    const std::shared_ptr<Kernel::Scheduler>& Scheduler(std::size_t core_index);
 
-    /// Gets the current process
-    Kernel::SharedPtr<Kernel::Process>& CurrentProcess() {
-        return current_process;
-    }
+    /// Provides a reference to the current process
+    Kernel::SharedPtr<Kernel::Process>& CurrentProcess();
 
-    PerfStats perf_stats;
-    FrameLimiter frame_limiter;
+    /// Provides a constant reference to the current process.
+    const Kernel::SharedPtr<Kernel::Process>& CurrentProcess() const;
 
-    void SetStatus(ResultStatus new_status, const char* details = nullptr) {
-        status = new_status;
-        if (details) {
-            status_details = details;
-        }
-    }
+    /// Provides a reference to the kernel instance.
+    Kernel::KernelCore& Kernel();
 
-    const std::string& GetStatusDetails() const {
-        return status_details;
-    }
+    /// Provides a constant reference to the kernel instance.
+    const Kernel::KernelCore& Kernel() const;
 
-    Loader::AppLoader& GetAppLoader() const {
-        return *app_loader;
-    }
+    /// Provides a reference to the internal PerfStats instance.
+    Core::PerfStats& GetPerfStats();
+
+    /// Provides a constant reference to the internal PerfStats instance.
+    const Core::PerfStats& GetPerfStats() const;
+
+    /// Provides a reference to the frame limiter;
+    Core::FrameLimiter& FrameLimiter();
+
+    /// Provides a constant referent to the frame limiter
+    const Core::FrameLimiter& FrameLimiter() const;
+
+    /// Gets the name of the current game
+    Loader::ResultStatus GetGameName(std::string& out) const;
+
+    void SetStatus(ResultStatus new_status, const char* details);
+
+    const std::string& GetStatusDetails() const;
+
+    Loader::AppLoader& GetAppLoader() const;
 
     Service::SM::ServiceManager& ServiceManager();
     const Service::SM::ServiceManager& ServiceManager() const;
 
-    void SetGPUDebugContext(std::shared_ptr<Tegra::DebugContext> context) {
-        debug_context = std::move(context);
-    }
+    void SetGPUDebugContext(std::shared_ptr<Tegra::DebugContext> context);
 
-    std::shared_ptr<Tegra::DebugContext> GetGPUDebugContext() const {
-        return debug_context;
-    }
+    Tegra::DebugContext* GetGPUDebugContext() const;
 
-    void SetFilesystem(FileSys::VirtualFilesystem vfs) {
-        virtual_filesystem = std::move(vfs);
-    }
+    void SetFilesystem(std::shared_ptr<FileSys::VfsFilesystem> vfs);
 
-    FileSys::VirtualFilesystem GetFilesystem() const {
-        return virtual_filesystem;
-    }
+    std::shared_ptr<FileSys::VfsFilesystem> GetFilesystem() const;
 
 private:
     System();
@@ -238,33 +232,10 @@ private:
      */
     ResultStatus Init(Frontend::EmuWindow& emu_window);
 
-    /// RealVfsFilesystem instance
-    FileSys::VirtualFilesystem virtual_filesystem;
-    /// AppLoader used to load the current executing application
-    std::unique_ptr<Loader::AppLoader> app_loader;
-    std::unique_ptr<VideoCore::RendererBase> renderer;
-    std::unique_ptr<Tegra::GPU> gpu_core;
-    std::shared_ptr<Tegra::DebugContext> debug_context;
-    Kernel::SharedPtr<Kernel::Process> current_process;
-    std::shared_ptr<ExclusiveMonitor> cpu_exclusive_monitor;
-    std::shared_ptr<CpuBarrier> cpu_barrier;
-    std::array<std::shared_ptr<Cpu>, NUM_CPU_CORES> cpu_cores;
-    std::array<std::unique_ptr<std::thread>, NUM_CPU_CORES - 1> cpu_core_threads;
-    size_t active_core{}; ///< Active core, only used in single thread mode
-
-    /// Service manager
-    std::shared_ptr<Service::SM::ServiceManager> service_manager;
-
-    /// Telemetry session for this emulation session
-    std::unique_ptr<Core::TelemetrySession> telemetry_session;
+    struct Impl;
+    std::unique_ptr<Impl> impl;
 
     static System s_instance;
-
-    ResultStatus status = ResultStatus::Success;
-    std::string status_details = "";
-
-    /// Map of guest threads to CPU cores
-    std::map<std::thread::id, std::shared_ptr<Cpu>> thread_to_cpu;
 };
 
 inline ARM_Interface& CurrentArmInterface() {

src/core/core_cpu.cpp

@@ -9,6 +9,7 @@
 #ifdef ARCHITECTURE_x86_64
 #include "core/arm/dynarmic/arm_dynarmic.h"
 #endif
+#include "core/arm/exclusive_monitor.h"
 #include "core/arm/unicorn/arm_unicorn.h"
 #include "core/core_cpu.h"
 #include "core/core_timing.h"
@@ -49,7 +50,7 @@ bool CpuBarrier::Rendezvous() {
 }
 
 Cpu::Cpu(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
-         std::shared_ptr<CpuBarrier> cpu_barrier, size_t core_index)
+         std::shared_ptr<CpuBarrier> cpu_barrier, std::size_t core_index)
     : cpu_barrier{std::move(cpu_barrier)}, core_index{core_index} {
 
     if (Settings::values.use_cpu_jit) {
@@ -66,7 +67,9 @@ Cpu::Cpu(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
     scheduler = std::make_shared<Kernel::Scheduler>(arm_interface.get());
 }
 
-std::shared_ptr<ExclusiveMonitor> Cpu::MakeExclusiveMonitor(size_t num_cores) {
+Cpu::~Cpu() = default;
+
+std::shared_ptr<ExclusiveMonitor> Cpu::MakeExclusiveMonitor(std::size_t num_cores) {
     if (Settings::values.use_cpu_jit) {
 #ifdef ARCHITECTURE_x86_64
         return std::make_shared<DynarmicExclusiveMonitor>(num_cores);

src/core/core_cpu.h

@@ -6,13 +6,10 @@
 
 #include <atomic>
 #include <condition_variable>
+#include <cstddef>
 #include <memory>
 #include <mutex>
-#include <string>
 #include "common/common_types.h"
-#include "core/arm/exclusive_monitor.h"
-
-class ARM_Interface;
 
 namespace Kernel {
 class Scheduler;
@@ -20,6 +17,9 @@ class Scheduler;
 
 namespace Core {
 
+class ARM_Interface;
+class ExclusiveMonitor;
+
 constexpr unsigned NUM_CPU_CORES{4};
 
 class CpuBarrier {
@@ -42,7 +42,8 @@ private:
 class Cpu {
 public:
     Cpu(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
-        std::shared_ptr<CpuBarrier> cpu_barrier, size_t core_index);
+        std::shared_ptr<CpuBarrier> cpu_barrier, std::size_t core_index);
+    ~Cpu();
 
     void RunLoop(bool tight_loop = true);
 
@@ -66,11 +67,11 @@ public:
         return core_index == 0;
     }
 
-    size_t CoreIndex() const {
+    std::size_t CoreIndex() const {
         return core_index;
     }
 
-    static std::shared_ptr<ExclusiveMonitor> MakeExclusiveMonitor(size_t num_cores);
+    static std::shared_ptr<ExclusiveMonitor> MakeExclusiveMonitor(std::size_t num_cores);
 
 private:
     void Reschedule();
@@ -80,7 +81,7 @@ private:
     std::shared_ptr<Kernel::Scheduler> scheduler;
 
     std::atomic<bool> reschedule_pending = false;
-    size_t core_index;
+    std::size_t core_index;
 };
 
 } // namespace Core

src/core/crypto/aes_util.cpp

@@ -10,9 +10,9 @@
 
 namespace Core::Crypto {
 namespace {
-std::vector<u8> CalculateNintendoTweak(size_t sector_id) {
+std::vector<u8> CalculateNintendoTweak(std::size_t sector_id) {
     std::vector<u8> out(0x10);
-    for (size_t i = 0xF; i <= 0xF; --i) {
+    for (std::size_t i = 0xF; i <= 0xF; --i) {
         out[i] = sector_id & 0xFF;
         sector_id >>= 8;
     }
@@ -20,11 +20,14 @@ std::vector<u8> CalculateNintendoTweak(size_t sector_id) {
 }
 } // Anonymous namespace
 
-static_assert(static_cast<size_t>(Mode::CTR) == static_cast<size_t>(MBEDTLS_CIPHER_AES_128_CTR),
+static_assert(static_cast<std::size_t>(Mode::CTR) ==
+                  static_cast<std::size_t>(MBEDTLS_CIPHER_AES_128_CTR),
               "CTR has incorrect value.");
-static_assert(static_cast<size_t>(Mode::ECB) == static_cast<size_t>(MBEDTLS_CIPHER_AES_128_ECB),
+static_assert(static_cast<std::size_t>(Mode::ECB) ==
+                  static_cast<std::size_t>(MBEDTLS_CIPHER_AES_128_ECB),
               "ECB has incorrect value.");
-static_assert(static_cast<size_t>(Mode::XTS) == static_cast<size_t>(MBEDTLS_CIPHER_AES_128_XTS),
+static_assert(static_cast<std::size_t>(Mode::XTS) ==
+                  static_cast<std::size_t>(MBEDTLS_CIPHER_AES_128_XTS),
               "XTS has incorrect value.");
 
 // Structure to hide mbedtls types from header file
@@ -33,7 +36,7 @@ struct CipherContext {
     mbedtls_cipher_context_t decryption_context;
 };
 
-template <typename Key, size_t KeySize>
+template <typename Key, std::size_t KeySize>
 Crypto::AESCipher<Key, KeySize>::AESCipher(Key key, Mode mode)
     : ctx(std::make_unique<CipherContext>()) {
     mbedtls_cipher_init(&ctx->encryption_context);
@@ -54,26 +57,26 @@ Crypto::AESCipher<Key, KeySize>::AESCipher(Key key, Mode mode)
     //"Failed to set key on mbedtls ciphers.");
 }
 
-template <typename Key, size_t KeySize>
+template <typename Key, std::size_t KeySize>
 AESCipher<Key, KeySize>::~AESCipher() {
     mbedtls_cipher_free(&ctx->encryption_context);
     mbedtls_cipher_free(&ctx->decryption_context);
 }
 
-template <typename Key, size_t KeySize>
+template <typename Key, std::size_t KeySize>
 void AESCipher<Key, KeySize>::SetIV(std::vector<u8> iv) {
     ASSERT_MSG((mbedtls_cipher_set_iv(&ctx->encryption_context, iv.data(), iv.size()) ||
                 mbedtls_cipher_set_iv(&ctx->decryption_context, iv.data(), iv.size())) == 0,
                "Failed to set IV on mbedtls ciphers.");
 }
 
-template <typename Key, size_t KeySize>
-void AESCipher<Key, KeySize>::Transcode(const u8* src, size_t size, u8* dest, Op op) const {
+template <typename Key, std::size_t KeySize>
+void AESCipher<Key, KeySize>::Transcode(const u8* src, std::size_t size, u8* dest, Op op) const {
     auto* const context = op == Op::Encrypt ? &ctx->encryption_context : &ctx->decryption_context;
 
     mbedtls_cipher_reset(context);
 
-    size_t written = 0;
+    std::size_t written = 0;
     if (mbedtls_cipher_get_cipher_mode(context) == MBEDTLS_MODE_XTS) {
         mbedtls_cipher_update(context, src, size, dest, &written);
         if (written != size) {
@@ -82,11 +85,25 @@ void AESCipher<Key, KeySize>::Transcode(const u8* src, size_t size, u8* dest, Op
         }
     } else {
         const auto block_size = mbedtls_cipher_get_block_size(context);
+        if (size < block_size) {
+            std::vector<u8> block(block_size);
+            std::memcpy(block.data(), src, size);
+            Transcode(block.data(), block.size(), block.data(), op);
+            std::memcpy(dest, block.data(), size);
+            return;
+        }
 
-        for (size_t offset = 0; offset < size; offset += block_size) {
-            auto length = std::min<size_t>(block_size, size - offset);
+        for (std::size_t offset = 0; offset < size; offset += block_size) {
+            auto length = std::min<std::size_t>(block_size, size - offset);
             mbedtls_cipher_update(context, src + offset, length, dest + offset, &written);
             if (written != length) {
+                if (length < block_size) {
+                    std::vector<u8> block(block_size);
+                    std::memcpy(block.data(), src + offset, length);
+                    Transcode(block.data(), block.size(), block.data(), op);
+                    std::memcpy(dest + offset, block.data(), length);
+                    return;
+                }
                 LOG_WARNING(Crypto, "Not all data was decrypted requested={:016X}, actual={:016X}.",
                             length, written);
             }
@@ -96,15 +113,12 @@ void AESCipher<Key, KeySize>::Transcode(const u8* src, size_t size, u8* dest, Op
     mbedtls_cipher_finish(context, nullptr, nullptr);
 }
 
-template <typename Key, size_t KeySize>
-void AESCipher<Key, KeySize>::XTSTranscode(const u8* src, size_t size, u8* dest, size_t sector_id,
-                                           size_t sector_size, Op op) {
-    if (size % sector_size > 0) {
-        LOG_CRITICAL(Crypto, "Data size must be a multiple of sector size.");
-        return;
-    }
+template <typename Key, std::size_t KeySize>
+void AESCipher<Key, KeySize>::XTSTranscode(const u8* src, std::size_t size, u8* dest,
+                                           std::size_t sector_id, std::size_t sector_size, Op op) {
+    ASSERT_MSG(size % sector_size == 0, "XTS decryption size must be a multiple of sector size.");
 
-    for (size_t i = 0; i < size; i += sector_size) {
+    for (std::size_t i = 0; i < size; i += sector_size) {
         SetIV(CalculateNintendoTweak(sector_id++));
         Transcode<u8, u8>(src + i, sector_size, dest + i, op);
     }
@@ -112,4 +126,4 @@ void AESCipher<Key, KeySize>::XTSTranscode(const u8* src, size_t size, u8* dest,
 
 template class AESCipher<Key128>;
 template class AESCipher<Key256>;
-} // namespace Core::Crypto
+} // namespace Core::Crypto

src/core/crypto/aes_util.h

@@ -25,7 +25,7 @@ enum class Op {
     Decrypt,
 };
 
-template <typename Key, size_t KeySize = sizeof(Key)>
+template <typename Key, std::size_t KeySize = sizeof(Key)>
 class AESCipher {
     static_assert(std::is_same_v<Key, std::array<u8, KeySize>>, "Key must be std::array of u8.");
     static_assert(KeySize == 0x10 || KeySize == 0x20, "KeySize must be 128 or 256.");
@@ -38,25 +38,25 @@ public:
     void SetIV(std::vector<u8> iv);
 
     template <typename Source, typename Dest>
-    void Transcode(const Source* src, size_t size, Dest* dest, Op op) const {
+    void Transcode(const Source* src, std::size_t size, Dest* dest, Op op) const {
         static_assert(std::is_trivially_copyable_v<Source> && std::is_trivially_copyable_v<Dest>,
                       "Transcode source and destination types must be trivially copyable.");
         Transcode(reinterpret_cast<const u8*>(src), size, reinterpret_cast<u8*>(dest), op);
     }
 
-    void Transcode(const u8* src, size_t size, u8* dest, Op op) const;
+    void Transcode(const u8* src, std::size_t size, u8* dest, Op op) const;
 
     template <typename Source, typename Dest>
-    void XTSTranscode(const Source* src, size_t size, Dest* dest, size_t sector_id,
-                      size_t sector_size, Op op) {
+    void XTSTranscode(const Source* src, std::size_t size, Dest* dest, std::size_t sector_id,
+                      std::size_t sector_size, Op op) {
         static_assert(std::is_trivially_copyable_v<Source> && std::is_trivially_copyable_v<Dest>,
                       "XTSTranscode source and destination types must be trivially copyable.");
         XTSTranscode(reinterpret_cast<const u8*>(src), size, reinterpret_cast<u8*>(dest), sector_id,
                      sector_size, op);
     }
 
-    void XTSTranscode(const u8* src, size_t size, u8* dest, size_t sector_id, size_t sector_size,
-                      Op op);
+    void XTSTranscode(const u8* src, std::size_t size, u8* dest, std::size_t sector_id,
+                      std::size_t sector_size, Op op);
 
 private:
     std::unique_ptr<CipherContext> ctx;

src/core/crypto/ctr_encryption_layer.cpp

@@ -8,11 +8,12 @@
 
 namespace Core::Crypto {
 
-CTREncryptionLayer::CTREncryptionLayer(FileSys::VirtualFile base_, Key128 key_, size_t base_offset)
+CTREncryptionLayer::CTREncryptionLayer(FileSys::VirtualFile base_, Key128 key_,
+                                       std::size_t base_offset)
     : EncryptionLayer(std::move(base_)), base_offset(base_offset), cipher(key_, Mode::CTR),
       iv(16, 0) {}
 
-size_t CTREncryptionLayer::Read(u8* data, size_t length, size_t offset) const {
+std::size_t CTREncryptionLayer::Read(u8* data, std::size_t length, std::size_t offset) const {
     if (length == 0)
         return 0;
 
@@ -20,9 +21,7 @@ size_t CTREncryptionLayer::Read(u8* data, size_t length, size_t offset) const {
     if (sector_offset == 0) {
         UpdateIV(base_offset + offset);
         std::vector<u8> raw = base->ReadBytes(length, offset);
-        if (raw.size() != length)
-            return Read(data, raw.size(), offset);
-        cipher.Transcode(raw.data(), length, data, Op::Decrypt);
+        cipher.Transcode(raw.data(), raw.size(), data, Op::Decrypt);
         return length;
     }
 
@@ -30,11 +29,11 @@ size_t CTREncryptionLayer::Read(u8* data, size_t length, size_t offset) const {
     std::vector<u8> block = base->ReadBytes(0x10, offset - sector_offset);
     UpdateIV(base_offset + offset - sector_offset);
     cipher.Transcode(block.data(), block.size(), block.data(), Op::Decrypt);
-    size_t read = 0x10 - sector_offset;
+    std::size_t read = 0x10 - sector_offset;
 
     if (length + sector_offset < 0x10) {
         std::memcpy(data, block.data() + sector_offset, std::min<u64>(length, read));
-        return read;
+        return std::min<u64>(length, read);
     }
     std::memcpy(data, block.data() + sector_offset, read);
     return read + Read(data + read, length - read, offset + read);
@@ -45,9 +44,9 @@ void CTREncryptionLayer::SetIV(const std::vector<u8>& iv_) {
     iv.assign(iv_.cbegin(), iv_.cbegin() + length);
 }
 
-void CTREncryptionLayer::UpdateIV(size_t offset) const {
+void CTREncryptionLayer::UpdateIV(std::size_t offset) const {
     offset >>= 4;
-    for (size_t i = 0; i < 8; ++i) {
+    for (std::size_t i = 0; i < 8; ++i) {
         iv[16 - i - 1] = offset & 0xFF;
         offset >>= 8;
     }

src/core/crypto/ctr_encryption_layer.h

@@ -14,20 +14,20 @@ namespace Core::Crypto {
 // Sits on top of a VirtualFile and provides CTR-mode AES decription.
 class CTREncryptionLayer : public EncryptionLayer {
 public:
-    CTREncryptionLayer(FileSys::VirtualFile base, Key128 key, size_t base_offset);
+    CTREncryptionLayer(FileSys::VirtualFile base, Key128 key, std::size_t base_offset);
 
-    size_t Read(u8* data, size_t length, size_t offset) const override;
+    std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override;
 
     void SetIV(const std::vector<u8>& iv);
 
 private:
-    size_t base_offset;
+    std::size_t base_offset;
 
     // Must be mutable as operations modify cipher contexts.
     mutable AESCipher<Key128> cipher;
     mutable std::vector<u8> iv;
 
-    void UpdateIV(size_t offset) const;
+    void UpdateIV(std::size_t offset) const;
 };
 
 } // namespace Core::Crypto

src/core/crypto/encryption_layer.cpp

@@ -12,11 +12,11 @@ std::string EncryptionLayer::GetName() const {
     return base->GetName();
 }
 
-size_t EncryptionLayer::GetSize() const {
+std::size_t EncryptionLayer::GetSize() const {
     return base->GetSize();
 }
 
-bool EncryptionLayer::Resize(size_t new_size) {
+bool EncryptionLayer::Resize(std::size_t new_size) {
     return false;
 }
 
@@ -32,7 +32,7 @@ bool EncryptionLayer::IsReadable() const {
     return true;
 }
 
-size_t EncryptionLayer::Write(const u8* data, size_t length, size_t offset) {
+std::size_t EncryptionLayer::Write(const u8* data, std::size_t length, std::size_t offset) {
     return 0;
 }
 

src/core/crypto/encryption_layer.h

@@ -15,15 +15,15 @@ class EncryptionLayer : public FileSys::VfsFile {
 public:
     explicit EncryptionLayer(FileSys::VirtualFile base);
 
-    size_t Read(u8* data, size_t length, size_t offset) const override = 0;
+    std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override = 0;
 
     std::string GetName() const override;
-    size_t GetSize() const override;
-    bool Resize(size_t new_size) override;
+    std::size_t GetSize() const override;
+    bool Resize(std::size_t new_size) override;
     std::shared_ptr<FileSys::VfsDirectory> GetContainingDirectory() const override;
     bool IsWritable() const override;
     bool IsReadable() const override;
-    size_t Write(const u8* data, size_t length, size_t offset) override;
+    std::size_t Write(const u8* data, std::size_t length, std::size_t offset) override;
     bool Rename(std::string_view name) override;
 
 protected: