nvidia_flags: Add missing header guard

Prevents potential inclusion compilation errors.

.ci/scripts/clang/docker.sh

@@ -0,0 +1,18 @@
+#!/bin/bash -ex
+
+# Exit on error, rather than continuing with the rest of the script.
+set -e
+
+cd /yuzu
+
+ccache -s
+
+mkdir build || true && cd build
+cmake .. -DDISPLAY_VERSION=$1 -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=/usr/lib/ccache/clang -DCMAKE_CXX_COMPILER=/usr/lib/ccache/clang++ -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${ENABLE_COMPATIBILITY_REPORTING:-"OFF"} -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DUSE_DISCORD_PRESENCE=ON -DENABLE_QT_TRANSLATION=ON -DCMAKE_INSTALL_PREFIX="/usr"
+
+make -j$(nproc)
+
+ccache -s
+
+ctest -VV -C Release
+

.ci/scripts/clang/exec.sh

@@ -0,0 +1,8 @@
+#!/bin/bash -ex
+
+mkdir -p "ccache"  || true
+chmod a+x ./.ci/scripts/clang/docker.sh
+# the UID for the container yuzu user is 1027
+sudo chown -R 1027 ./
+docker run -e ENABLE_COMPATIBILITY_REPORTING -e CCACHE_DIR=/yuzu/ccache -v $(pwd):/yuzu yuzuemu/build-environments:linux-fresh /bin/bash /yuzu/.ci/scripts/clang/docker.sh $1
+sudo chown -R $UID ./

.ci/scripts/clang/upload.sh

@@ -0,0 +1,20 @@
+#!/bin/bash -ex
+
+. .ci/scripts/common/pre-upload.sh
+
+REV_NAME="yuzu-linux-${GITDATE}-${GITREV}"
+ARCHIVE_NAME="${REV_NAME}.tar.xz"
+COMPRESSION_FLAGS="-cJvf"
+
+if [ "${RELEASE_NAME}" = "mainline" ]; then
+    DIR_NAME="${REV_NAME}"
+else
+    DIR_NAME="${REV_NAME}_${RELEASE_NAME}"
+fi
+
+mkdir "$DIR_NAME"
+
+cp build/bin/yuzu-cmd "$DIR_NAME"
+cp build/bin/yuzu "$DIR_NAME"
+
+. .ci/scripts/common/post-upload.sh

.ci/templates/build-msvc.yml

@@ -8,7 +8,7 @@ steps:
   displayName: 'Install vulkan-sdk'
 - script: python -m pip install --upgrade pip conan
   displayName: 'Install conan'
-- script: refreshenv && mkdir build && cd build && cmake -G "Visual Studio 16 2019" -A x64 --config Release -DYUZU_USE_BUNDLED_QT=1 -DYUZU_USE_QT_WEB_ENGINE=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${COMPAT} -DUSE_DISCORD_PRESENCE=ON -DENABLE_QT_TRANSLATION=ON -DDISPLAY_VERSION=${{ parameters['version'] }} .. && cd ..
+- script: refreshenv && mkdir build && cd build && cmake -G "Visual Studio 16 2019" -A x64 -DYUZU_USE_BUNDLED_QT=1 -DYUZU_USE_QT_WEB_ENGINE=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${COMPAT} -DUSE_DISCORD_PRESENCE=ON -DENABLE_QT_TRANSLATION=ON -DDISPLAY_VERSION=${{ parameters['version'] }} .. && cmake --install . --config Release && cd ..
   displayName: 'Configure CMake'
 - task: MSBuild@1
   displayName: 'Build'

.ci/templates/build-standard.yml

@@ -12,6 +12,9 @@ jobs:
       windows:
         BuildSuffix: 'windows-mingw'
         ScriptFolder: 'windows'
+      clang:
+        BuildSuffix: 'clang'
+        ScriptFolder: 'clang'
       linux:
         BuildSuffix: 'linux'
         ScriptFolder: 'linux'
@@ -24,4 +27,4 @@ jobs:
     parameters:
       artifactSource: 'false'
       cache: $(parameters.cache)
-      version: $(parameters.version)
+      version: $(parameters.version)

.gitmodules

@@ -27,7 +27,7 @@
     url = https://github.com/ReinUsesLisp/sirit
 [submodule "mbedtls"]
     path = externals/mbedtls
-    url = https://github.com/DarkLordZach/mbedtls
+    url = https://github.com/yuzu-emu/mbedtls
 [submodule "libzip"]
     path = externals/libzip/libzip
     url = https://github.com/nih-at/libzip.git
@@ -37,6 +37,6 @@
 [submodule "opus"]
 	path = externals/opus/opus
 	url = https://github.com/xiph/opus.git
-[submodule "externals/ffmpeg"]
+[submodule "ffmpeg"]
 	path = externals/ffmpeg
 	url = https://git.ffmpeg.org/ffmpeg.git

CMakeLists.txt

@@ -172,6 +172,8 @@ macro(yuzu_find_packages)
         "nlohmann_json     3.8         nlohmann_json/3.8.0"
         "ZLIB              1.2         zlib/1.2.11"
         "zstd              1.4         zstd/1.4.8"
+    # can't use opus until AVX check is fixed: https://github.com/yuzu-emu/yuzu/pull/4068
+        #"opus              1.3         opus/1.3.1"
     )
 
     foreach(PACKAGE ${REQUIRED_LIBS})
@@ -504,7 +506,7 @@ if (YUZU_USE_BUNDLED_FFMPEG)
         endif()
     else() # WIN32
         # Use yuzu FFmpeg binaries
-        set(FFmpeg_EXT_NAME "ffmpeg-4.2.1")
+        set(FFmpeg_EXT_NAME "ffmpeg-4.3.1")
         set(FFmpeg_PATH "${CMAKE_BINARY_DIR}/externals/${FFmpeg_EXT_NAME}")
         download_bundled_external("ffmpeg/" ${FFmpeg_EXT_NAME} "")
         set(FFmpeg_FOUND YES)

CMakeModules/CopyYuzuFFmpegDeps.cmake

@@ -1,10 +1,6 @@
 function(copy_yuzu_FFmpeg_deps target_dir)
     include(WindowsCopyFiles)
     set(DLL_DEST "${CMAKE_BINARY_DIR}/bin/$<CONFIG>/")
-    windows_copy_files(${target_dir} ${FFmpeg_DLL_DIR} ${DLL_DEST}
-        avcodec-58.dll
-        avutil-56.dll
-        swresample-3.dll
-        swscale-5.dll
-    )
+    file(READ "${FFmpeg_PATH}/requirements.txt" FFmpeg_REQUIRED_DLLS)
+    windows_copy_files(${target_dir} ${FFmpeg_DLL_DIR} ${DLL_DEST} ${FFmpeg_REQUIRED_DLLS})
 endfunction(copy_yuzu_FFmpeg_deps)

dist/qt_themes/colorful_dark/icons/index.theme

@@ -1,7 +1,7 @@
 [Icon Theme]
 Name=colorful_dark
 Comment=Colorful theme (Dark style)
-Inherits=default
+Inherits=colorful
 Directories=16x16
  
 [16x16]

dist/qt_themes/colorful_midnight_blue/icons/index.theme

@@ -1,7 +1,7 @@
 [Icon Theme]
 Name=colorful_midnight_blue
 Comment=Colorful theme (Midnight Blue style)
-Inherits=default
+Inherits=colorful
 Directories=16x16
 
 [16x16]

dist/qt_themes/qdarkstyle_midnight_blue/style.qss

@@ -1257,10 +1257,6 @@ QComboBox::item:alternate {
   background: #19232D;
 }
 
-QComboBox::item:checked {
-  font-weight: bold;
-}
-
 QComboBox::item:selected {
   border: 0px solid transparent;
 }

externals/CMakeLists.txt

@@ -64,8 +64,8 @@ endif()
 add_subdirectory(sirit)
 
 # libzip
-find_package(Libzip 1.5)
-if (NOT LIBZIP_FOUND)
+find_package(libzip 1.5)
+if (NOT libzip_FOUND)
     message(STATUS "libzip 1.5 or newer not found, falling back to externals")
     add_subdirectory(libzip EXCLUDE_FROM_ALL)
 endif()
@@ -97,4 +97,8 @@ if (ENABLE_WEB_SERVICE)
 endif()
 
 # Opus
-add_subdirectory(opus)
+find_package(opus 1.3)
+if (NOT opus_FOUND)
+    message(STATUS "opus 1.3 or newer not found, falling back to externals")
+    add_subdirectory(opus EXCLUDE_FROM_ALL)
+endif()

externals/find-modules/FindLibzip.cmake

@@ -1,72 +0,0 @@
-
-find_package(PkgConfig QUIET)
-pkg_check_modules(PC_LIBZIP QUIET libzip)
-
-find_path(LIBZIP_INCLUDE_DIR
-  NAMES zip.h
-  PATHS ${PC_LIBZIP_INCLUDE_DIRS}
-  "$ENV{LIB_DIR}/include"
-  "$ENV{INCLUDE}"
-  /usr/local/include
-  /usr/include
-)
-find_path(LIBZIP_INCLUDE_DIR_ZIPCONF
-  NAMES zipconf.h
-  HINTS ${PC_LIBZIP_INCLUDE_DIRS}
-  "$ENV{LIB_DIR}/include"
-  "$ENV{LIB_DIR}/lib/libzip/include"
-  "$ENV{LIB}/lib/libzip/include"
-  /usr/local/lib/libzip/include
-  /usr/lib/libzip/include
-  /usr/local/include
-  /usr/include
-  "$ENV{INCLUDE}"
-)
-find_library(LIBZIP_LIBRARY
-  NAMES zip
-  PATHS ${PC_LIBZIP_LIBRARY_DIRS}
-  "$ENV{LIB_DIR}/lib" "$ENV{LIB}" /usr/local/lib /usr/lib
-)
-
-if (LIBZIP_INCLUDE_DIR_ZIPCONF)
-  FILE(READ "${LIBZIP_INCLUDE_DIR_ZIPCONF}/zipconf.h" _LIBZIP_VERSION_CONTENTS)
-  if (_LIBZIP_VERSION_CONTENTS)
-    STRING(REGEX REPLACE ".*#define LIBZIP_VERSION \"([0-9.]+)\".*" "\\1" LIBZIP_VERSION "${_LIBZIP_VERSION_CONTENTS}")
-  endif()
-  unset(_LIBZIP_VERSION_CONTENTS)
-endif()
-
-set(LIBZIP_VERSION ${LIBZIP_VERSION} CACHE STRING "Version number of libzip")
-
-include(FindPackageHandleStandardArgs)
-find_package_handle_standard_args(Libzip
-  FOUND_VAR LIBZIP_FOUND
-  REQUIRED_VARS
-    LIBZIP_LIBRARY
-    LIBZIP_INCLUDE_DIR
-    LIBZIP_INCLUDE_DIR_ZIPCONF
-    LIBZIP_VERSION
-  VERSION_VAR LIBZIP_VERSION
-)
-
-if(LIBZIP_FOUND)
-  set(LIBZIP_LIBRARIES ${LIBZIP_LIBRARY})
-  set(LIBZIP_INCLUDE_DIRS ${LIBZIP_INCLUDE_DIR})
-  set(LIBZIP_DEFINITIONS ${PC_LIBZIP_CFLAGS_OTHER})
-endif()
-
-if(LIBZIP_FOUND AND NOT TARGET libzip::libzip)
-  add_library(libzip::libzip UNKNOWN IMPORTED)
-  set_target_properties(libzip::libzip PROPERTIES
-    IMPORTED_LOCATION "${LIBZIP_LIBRARY}"
-    INTERFACE_COMPILE_OPTIONS "${PC_LIBZIP_CFLAGS_OTHER}"
-    INTERFACE_INCLUDE_DIRECTORIES "${LIBZIP_INCLUDE_DIR}"
-  )
-endif()
-
-mark_as_advanced(
-    LIBZIP_INCLUDE_DIR
-    LIBZIP_INCLUDE_DIR_ZIPCONF
-    LIBZIP_LIBRARY
-    LIBZIP_VERSION
-)

externals/find-modules/Findlibzip.cmake

@@ -0,0 +1,72 @@
+
+find_package(PkgConfig QUIET)
+pkg_check_modules(PC_libzip QUIET libzip)
+
+find_path(libzip_INCLUDE_DIR
+  NAMES zip.h
+  PATHS ${PC_libzip_INCLUDE_DIRS}
+  "$ENV{LIB_DIR}/include"
+  "$ENV{INCLUDE}"
+  /usr/local/include
+  /usr/include
+)
+find_path(libzip_INCLUDE_DIR_ZIPCONF
+  NAMES zipconf.h
+  HINTS ${PC_libzip_INCLUDE_DIRS}
+  "$ENV{LIB_DIR}/include"
+  "$ENV{LIB_DIR}/lib/libzip/include"
+  "$ENV{LIB}/lib/libzip/include"
+  /usr/local/lib/libzip/include
+  /usr/lib/libzip/include
+  /usr/local/include
+  /usr/include
+  "$ENV{INCLUDE}"
+)
+find_library(libzip_LIBRARY
+  NAMES zip
+  PATHS ${PC_libzip_LIBRARY_DIRS}
+  "$ENV{LIB_DIR}/lib" "$ENV{LIB}" /usr/local/lib /usr/lib
+)
+
+if (libzip_INCLUDE_DIR_ZIPCONF)
+  FILE(READ "${libzip_INCLUDE_DIR_ZIPCONF}/zipconf.h" _libzip_VERSION_CONTENTS)
+  if (_libzip_VERSION_CONTENTS)
+    STRING(REGEX REPLACE ".*#define LIBZIP_VERSION \"([0-9.]+)\".*" "\\1" libzip_VERSION "${_libzip_VERSION_CONTENTS}")
+  endif()
+  unset(_libzip_VERSION_CONTENTS)
+endif()
+
+set(libzip_VERSION ${libzip_VERSION} CACHE STRING "Version number of libzip")
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(libzip
+  FOUND_VAR libzip_FOUND
+  REQUIRED_VARS
+    libzip_LIBRARY
+    libzip_INCLUDE_DIR
+    libzip_INCLUDE_DIR_ZIPCONF
+    libzip_VERSION
+  VERSION_VAR libzip_VERSION
+)
+
+if(libzip_FOUND)
+  set(libzip_LIBRARIES ${libzip_LIBRARY})
+  set(libzip_INCLUDE_DIRS ${libzip_INCLUDE_DIR})
+  set(libzip_DEFINITIONS ${PC_libzip_CFLAGS_OTHER})
+endif()
+
+if(libzip_FOUND AND NOT TARGET libzip::libzip)
+  add_library(libzip::libzip UNKNOWN IMPORTED)
+  set_target_properties(libzip::libzip PROPERTIES
+    IMPORTED_LOCATION "${libzip_LIBRARY}"
+    INTERFACE_COMPILE_OPTIONS "${PC_libzip_CFLAGS_OTHER}"
+    INTERFACE_INCLUDE_DIRECTORIES "${libzip_INCLUDE_DIR}"
+  )
+endif()
+
+mark_as_advanced(
+    libzip_INCLUDE_DIR
+    libzip_INCLUDE_DIR_ZIPCONF
+    libzip_LIBRARY
+    libzip_VERSION
+)

externals/find-modules/Findopus.cmake

@@ -28,7 +28,7 @@ if(opus_FOUND)
 endif()
 
 if(opus_FOUND AND NOT TARGET Opus::Opus)
-  add_library(Opus::Opus UNKNOWN IMPORTED)
+  add_library(Opus::Opus UNKNOWN IMPORTED GLOBAL)
   set_target_properties(Opus::Opus PROPERTIES
     IMPORTED_LOCATION "${opus_LIBRARY}"
     INTERFACE_COMPILE_OPTIONS "${PC_opus_CFLAGS_OTHER}"

externals/glad/include/glad/glad.h

@@ -5156,6 +5156,9 @@ GLAPI PFNGLDEPTHRANGEARRAYVPROC glad_glDepthRangeArrayv;
 typedef void (APIENTRYP PFNGLDEPTHRANGEINDEXEDPROC)(GLuint index, GLdouble n, GLdouble f);
 GLAPI PFNGLDEPTHRANGEINDEXEDPROC glad_glDepthRangeIndexed;
 #define glDepthRangeIndexed glad_glDepthRangeIndexed
+typedef void (APIENTRYP PFNGLDEPTHRANGEINDEXEDDNVPROC)(GLuint index, GLdouble n, GLdouble f);
+GLAPI PFNGLDEPTHRANGEINDEXEDDNVPROC glad_glDepthRangeIndexeddNV;
+#define glDepthRangeIndexeddNV glad_glDepthRangeIndexeddNV
 typedef void (APIENTRYP PFNGLGETFLOATI_VPROC)(GLenum target, GLuint index, GLfloat *data);
 GLAPI PFNGLGETFLOATI_VPROC glad_glGetFloati_v;
 #define glGetFloati_v glad_glGetFloati_v

externals/glad/src/glad.c

@@ -1044,6 +1044,7 @@ PFNGLDEPTHMASKPROC glad_glDepthMask = NULL;
 PFNGLDEPTHRANGEPROC glad_glDepthRange = NULL;
 PFNGLDEPTHRANGEARRAYVPROC glad_glDepthRangeArrayv = NULL;
 PFNGLDEPTHRANGEINDEXEDPROC glad_glDepthRangeIndexed = NULL;
+PFNGLDEPTHRANGEINDEXEDDNVPROC glad_glDepthRangeIndexeddNV = NULL;
 PFNGLDEPTHRANGEFPROC glad_glDepthRangef = NULL;
 PFNGLDETACHSHADERPROC glad_glDetachShader = NULL;
 PFNGLDISABLEPROC glad_glDisable = NULL;
@@ -7971,6 +7972,7 @@ static void load_GL_NV_depth_buffer_float(GLADloadproc load) {
 	glad_glDepthRangedNV = (PFNGLDEPTHRANGEDNVPROC)load("glDepthRangedNV");
 	glad_glClearDepthdNV = (PFNGLCLEARDEPTHDNVPROC)load("glClearDepthdNV");
 	glad_glDepthBoundsdNV = (PFNGLDEPTHBOUNDSDNVPROC)load("glDepthBoundsdNV");
+	glad_glDepthRangeIndexeddNV = (PFNGLDEPTHRANGEINDEXEDDNVPROC)load("glDepthRangeIndexeddNV");
 }
 static void load_GL_NV_draw_texture(GLADloadproc load) {
 	if(!GLAD_GL_NV_draw_texture) return;

externals/libusb/CMakeLists.txt

@@ -1,3 +1,8 @@
+# Ensure libusb compiles with UTF-8 encoding on MSVC
+if(MSVC)
+    add_compile_options(/utf-8)
+endif()
+
 add_library(usb STATIC EXCLUDE_FROM_ALL
     libusb/libusb/core.c
     libusb/libusb/core.c

externals/opus/CMakeLists.txt

@@ -252,3 +252,5 @@ PRIVATE
     opus/silk/float
     opus/src
 )
+
+add_library(Opus::Opus ALIAS opus)

src/CMakeLists.txt

@@ -27,6 +27,7 @@ if (MSVC)
     # /Zo                 - Enhanced debug info for optimized builds
     # /permissive-        - Enables stricter C++ standards conformance checks
     # /EHsc               - C++-only exception handling semantics
+    # /utf-8              - Set source and execution character sets to UTF-8
     # /volatile:iso       - Use strict standards-compliant volatile semantics.
     # /Zc:externConstexpr - Allow extern constexpr variables to have external linkage, like the standard mandates
     # /Zc:inline          - Let codegen omit inline functions in object files
@@ -38,6 +39,7 @@ if (MSVC)
         /permissive-
         /EHsc
         /std:c++latest
+        /utf-8
         /volatile:iso
         /Zc:externConstexpr
         /Zc:inline

src/audio_core/CMakeLists.txt

@@ -15,6 +15,8 @@ add_library(audio_core STATIC
     command_generator.cpp
     command_generator.h
     common.h
+    delay_line.cpp
+    delay_line.h
     effect_context.cpp
     effect_context.h
     info_updater.cpp

src/audio_core/command_generator.cpp

@@ -2,6 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cmath>
+#include <numbers>
 #include "audio_core/algorithm/interpolate.h"
 #include "audio_core/command_generator.h"
 #include "audio_core/effect_context.h"
@@ -13,6 +15,20 @@ namespace AudioCore {
 namespace {
 constexpr std::size_t MIX_BUFFER_SIZE = 0x3f00;
 constexpr std::size_t SCALED_MIX_BUFFER_SIZE = MIX_BUFFER_SIZE << 15ULL;
+using DelayLineTimes = std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT>;
+
+constexpr DelayLineTimes FDN_MIN_DELAY_LINE_TIMES{5.0f, 6.0f, 13.0f, 14.0f};
+constexpr DelayLineTimes FDN_MAX_DELAY_LINE_TIMES{45.704f, 82.782f, 149.94f, 271.58f};
+constexpr DelayLineTimes DECAY0_MAX_DELAY_LINE_TIMES{17.0f, 13.0f, 9.0f, 7.0f};
+constexpr DelayLineTimes DECAY1_MAX_DELAY_LINE_TIMES{19.0f, 11.0f, 10.0f, 6.0f};
+constexpr std::array<f32, AudioCommon::I3DL2REVERB_TAPS> EARLY_TAP_TIMES{
+    0.017136f, 0.059154f, 0.161733f, 0.390186f, 0.425262f, 0.455411f, 0.689737f,
+    0.745910f, 0.833844f, 0.859502f, 0.000000f, 0.075024f, 0.168788f, 0.299901f,
+    0.337443f, 0.371903f, 0.599011f, 0.716741f, 0.817859f, 0.851664f};
+constexpr std::array<f32, AudioCommon::I3DL2REVERB_TAPS> EARLY_GAIN{
+    0.67096f, 0.61027f, 1.0f,     0.35680f, 0.68361f, 0.65978f, 0.51939f,
+    0.24712f, 0.45945f, 0.45021f, 0.64196f, 0.54879f, 0.92925f, 0.38270f,
+    0.72867f, 0.69794f, 0.5464f,  0.24563f, 0.45214f, 0.44042f};
 
 template <std::size_t N>
 void ApplyMix(s32* output, const s32* input, s32 gain, s32 sample_count) {
@@ -65,6 +81,154 @@ s32 ApplyMixDepop(s32* output, s32 first_sample, s32 delta, s32 sample_count) {
     }
 }
 
+float Pow10(float x) {
+    if (x >= 0.0f) {
+        return 1.0f;
+    } else if (x <= -5.3f) {
+        return 0.0f;
+    }
+    return std::pow(10.0f, x);
+}
+
+float SinD(float degrees) {
+    return std::sin(degrees * std::numbers::pi_v<float> / 180.0f);
+}
+
+float CosD(float degrees) {
+    return std::cos(degrees * std::numbers::pi_v<float> / 180.0f);
+}
+
+float ToFloat(s32 sample) {
+    return static_cast<float>(sample) / 65536.f;
+}
+
+s32 ToS32(float sample) {
+    constexpr auto min = -8388608.0f;
+    constexpr auto max = 8388607.f;
+    float rescaled_sample = sample * 65536.0f;
+    if (rescaled_sample < min) {
+        rescaled_sample = min;
+    }
+    if (rescaled_sample > max) {
+        rescaled_sample = max;
+    }
+    return static_cast<s32>(rescaled_sample);
+}
+
+constexpr std::array<std::size_t, 20> REVERB_TAP_INDEX_1CH{0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                                                           0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
+
+constexpr std::array<std::size_t, 20> REVERB_TAP_INDEX_2CH{0, 0, 0, 1, 1, 1, 1, 0, 0, 0,
+                                                           1, 1, 1, 0, 0, 0, 0, 1, 1, 1};
+
+constexpr std::array<std::size_t, 20> REVERB_TAP_INDEX_4CH{0, 0, 0, 1, 1, 1, 1, 2, 2, 2,
+                                                           1, 1, 1, 0, 0, 0, 0, 3, 3, 3};
+
+constexpr std::array<std::size_t, 20> REVERB_TAP_INDEX_6CH{4, 0, 0, 1, 1, 1, 1, 2, 2, 2,
+                                                           1, 1, 1, 0, 0, 0, 0, 3, 3, 3};
+
+template <std::size_t CHANNEL_COUNT>
+void ApplyReverbGeneric(I3dl2ReverbState& state,
+                        const std::array<const s32*, AudioCommon::MAX_CHANNEL_COUNT>& input,
+                        const std::array<s32*, AudioCommon::MAX_CHANNEL_COUNT>& output,
+                        s32 sample_count) {
+
+    auto GetTapLookup = []() {
+        if constexpr (CHANNEL_COUNT == 1) {
+            return REVERB_TAP_INDEX_1CH;
+        } else if constexpr (CHANNEL_COUNT == 2) {
+            return REVERB_TAP_INDEX_2CH;
+        } else if constexpr (CHANNEL_COUNT == 4) {
+            return REVERB_TAP_INDEX_4CH;
+        } else if constexpr (CHANNEL_COUNT == 6) {
+            return REVERB_TAP_INDEX_6CH;
+        }
+    };
+
+    const auto& tap_index_lut = GetTapLookup();
+    for (s32 sample = 0; sample < sample_count; sample++) {
+        std::array<f32, CHANNEL_COUNT> out_samples{};
+        std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> fsamp{};
+        std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> mixed{};
+        std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> osamp{};
+
+        // Mix everything into a single sample
+        s32 temp_mixed_sample = 0;
+        for (std::size_t i = 0; i < CHANNEL_COUNT; i++) {
+            temp_mixed_sample += input[i][sample];
+        }
+        const auto current_sample = ToFloat(temp_mixed_sample);
+        const auto early_tap = state.early_delay_line.TapOut(state.early_to_late_taps);
+
+        for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_TAPS; i++) {
+            const auto tapped_samp =
+                state.early_delay_line.TapOut(state.early_tap_steps[i]) * EARLY_GAIN[i];
+            out_samples[tap_index_lut[i]] += tapped_samp;
+
+            if constexpr (CHANNEL_COUNT == 6) {
+                // handle lfe
+                out_samples[5] += tapped_samp;
+            }
+        }
+
+        state.lowpass_0 = current_sample * state.lowpass_2 + state.lowpass_0 * state.lowpass_1;
+        state.early_delay_line.Tick(state.lowpass_0);
+
+        for (std::size_t i = 0; i < CHANNEL_COUNT; i++) {
+            out_samples[i] *= state.early_gain;
+        }
+
+        // Two channel seems to apply a latet gain, we require to save this
+        f32 filter{};
+        for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+            filter = state.fdn_delay_line[i].GetOutputSample();
+            const auto computed = filter * state.lpf_coefficients[0][i] + state.shelf_filter[i];
+            state.shelf_filter[i] =
+                filter * state.lpf_coefficients[1][i] + computed * state.lpf_coefficients[2][i];
+            fsamp[i] = computed;
+        }
+
+        // Mixing matrix
+        mixed[0] = fsamp[1] + fsamp[2];
+        mixed[1] = -fsamp[0] - fsamp[3];
+        mixed[2] = fsamp[0] - fsamp[3];
+        mixed[3] = fsamp[1] - fsamp[2];
+
+        if constexpr (CHANNEL_COUNT == 2) {
+            for (auto& mix : mixed) {
+                mix *= (filter * state.late_gain);
+            }
+        }
+
+        for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+            const auto late = early_tap * state.late_gain;
+            osamp[i] = state.decay_delay_line0[i].Tick(late + mixed[i]);
+            osamp[i] = state.decay_delay_line1[i].Tick(osamp[i]);
+            state.fdn_delay_line[i].Tick(osamp[i]);
+        }
+
+        if constexpr (CHANNEL_COUNT == 1) {
+            output[0][sample] = ToS32(state.dry_gain * ToFloat(input[0][sample]) +
+                                      (out_samples[0] + osamp[0] + osamp[1]));
+        } else if constexpr (CHANNEL_COUNT == 2 || CHANNEL_COUNT == 4) {
+            for (std::size_t i = 0; i < CHANNEL_COUNT; i++) {
+                output[i][sample] =
+                    ToS32(state.dry_gain * ToFloat(input[i][sample]) + (out_samples[i] + osamp[i]));
+            }
+        } else if constexpr (CHANNEL_COUNT == 6) {
+            const auto temp_center = state.center_delay_line.Tick(0.5f * (osamp[2] - osamp[3]));
+            for (std::size_t i = 0; i < 4; i++) {
+                output[i][sample] =
+                    ToS32(state.dry_gain * ToFloat(input[i][sample]) + (out_samples[i] + osamp[i]));
+            }
+            output[4][sample] =
+                ToS32(state.dry_gain * ToFloat(input[4][sample]) + (out_samples[4] + temp_center));
+            output[5][sample] =
+                ToS32(state.dry_gain * ToFloat(input[5][sample]) + (out_samples[5] + osamp[3]));
+        }
+    }
+}
+
 } // namespace
 
 CommandGenerator::CommandGenerator(AudioCommon::AudioRendererParameter& worker_params_,
@@ -271,11 +435,10 @@ void CommandGenerator::GenerateBiquadFilterCommandForVoice(ServerVoiceInfo& voic
         }
 
         // Generate biquad filter
-        //        GenerateBiquadFilterCommand(mix_buffer_count, biquad_filter,
-        //        dsp_state.biquad_filter_state,
-        //                                    mix_buffer_count + channel, mix_buffer_count +
-        //                                    channel, worker_params.sample_count,
-        //                                    voice_info.GetInParams().node_id);
+        // GenerateBiquadFilterCommand(mix_buffer_count, biquad_filter,
+        // dsp_state.biquad_filter_state,
+        //                            mix_buffer_count + channel, mix_buffer_count + channel,
+        //                            worker_params.sample_count, voice_info.GetInParams().node_id);
     }
 }
 
@@ -376,21 +539,54 @@ void CommandGenerator::GenerateEffectCommand(ServerMixInfo& mix_info) {
 
 void CommandGenerator::GenerateI3dl2ReverbEffectCommand(s32 mix_buffer_offset, EffectBase* info,
                                                         bool enabled) {
-    if (!enabled) {
+    auto* reverb = dynamic_cast<EffectI3dl2Reverb*>(info);
+    const auto& params = reverb->GetParams();
+    auto& state = reverb->GetState();
+    const auto channel_count = params.channel_count;
+
+    if (channel_count != 1 && channel_count != 2 && channel_count != 4 && channel_count != 6) {
         return;
     }
-    const auto& params = dynamic_cast<EffectI3dl2Reverb*>(info)->GetParams();
-    const auto channel_count = params.channel_count;
+
+    std::array<const s32*, AudioCommon::MAX_CHANNEL_COUNT> input{};
+    std::array<s32*, AudioCommon::MAX_CHANNEL_COUNT> output{};
+
+    const auto status = params.status;
     for (s32 i = 0; i < channel_count; i++) {
-        // TODO(ogniK): Actually implement reverb
-        /*
-        if (params.input[i] != params.output[i]) {
-            const auto* input = GetMixBuffer(mix_buffer_offset + params.input[i]);
-            auto* output = GetMixBuffer(mix_buffer_offset + params.output[i]);
-            ApplyMix<1>(output, input, 32768, worker_params.sample_count);
-        }*/
-        auto* output = GetMixBuffer(mix_buffer_offset + params.output[i]);
-        std::memset(output, 0, worker_params.sample_count * sizeof(s32));
+        input[i] = GetMixBuffer(mix_buffer_offset + params.input[i]);
+        output[i] = GetMixBuffer(mix_buffer_offset + params.output[i]);
+    }
+
+    if (enabled) {
+        if (status == ParameterStatus::Initialized) {
+            InitializeI3dl2Reverb(reverb->GetParams(), state, info->GetWorkBuffer());
+        } else if (status == ParameterStatus::Updating) {
+            UpdateI3dl2Reverb(reverb->GetParams(), state, false);
+        }
+    }
+
+    if (enabled) {
+        switch (channel_count) {
+        case 1:
+            ApplyReverbGeneric<1>(state, input, output, worker_params.sample_count);
+            break;
+        case 2:
+            ApplyReverbGeneric<2>(state, input, output, worker_params.sample_count);
+            break;
+        case 4:
+            ApplyReverbGeneric<4>(state, input, output, worker_params.sample_count);
+            break;
+        case 6:
+            ApplyReverbGeneric<6>(state, input, output, worker_params.sample_count);
+            break;
+        }
+    } else {
+        for (s32 i = 0; i < channel_count; i++) {
+            // Only copy if the buffer input and output do not match!
+            if ((mix_buffer_offset + params.input[i]) != (mix_buffer_offset + params.output[i])) {
+                std::memcpy(output[i], input[i], worker_params.sample_count * sizeof(s32));
+            }
+        }
     }
 }
 
@@ -528,6 +724,133 @@ s32 CommandGenerator::ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u3
     return sample_count;
 }
 
+void CommandGenerator::InitializeI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state,
+                                             std::vector<u8>& work_buffer) {
+    // Reset state
+    state.lowpass_0 = 0.0f;
+    state.lowpass_1 = 0.0f;
+    state.lowpass_2 = 0.0f;
+
+    state.early_delay_line.Reset();
+    state.early_tap_steps.fill(0);
+    state.early_gain = 0.0f;
+    state.late_gain = 0.0f;
+    state.early_to_late_taps = 0;
+    for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+        state.fdn_delay_line[i].Reset();
+        state.decay_delay_line0[i].Reset();
+        state.decay_delay_line1[i].Reset();
+    }
+    state.last_reverb_echo = 0.0f;
+    state.center_delay_line.Reset();
+    for (auto& coef : state.lpf_coefficients) {
+        coef.fill(0.0f);
+    }
+    state.shelf_filter.fill(0.0f);
+    state.dry_gain = 0.0f;
+
+    const auto sample_rate = info.sample_rate / 1000;
+    f32* work_buffer_ptr = reinterpret_cast<f32*>(work_buffer.data());
+
+    s32 delay_samples{};
+    for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+        delay_samples =
+            AudioCommon::CalculateDelaySamples(sample_rate, FDN_MAX_DELAY_LINE_TIMES[i]);
+        state.fdn_delay_line[i].Initialize(delay_samples, work_buffer_ptr);
+        work_buffer_ptr += delay_samples + 1;
+
+        delay_samples =
+            AudioCommon::CalculateDelaySamples(sample_rate, DECAY0_MAX_DELAY_LINE_TIMES[i]);
+        state.decay_delay_line0[i].Initialize(delay_samples, 0.0f, work_buffer_ptr);
+        work_buffer_ptr += delay_samples + 1;
+
+        delay_samples =
+            AudioCommon::CalculateDelaySamples(sample_rate, DECAY1_MAX_DELAY_LINE_TIMES[i]);
+        state.decay_delay_line1[i].Initialize(delay_samples, 0.0f, work_buffer_ptr);
+        work_buffer_ptr += delay_samples + 1;
+    }
+    delay_samples = AudioCommon::CalculateDelaySamples(sample_rate, 5.0f);
+    state.center_delay_line.Initialize(delay_samples, work_buffer_ptr);
+    work_buffer_ptr += delay_samples + 1;
+
+    delay_samples = AudioCommon::CalculateDelaySamples(sample_rate, 400.0f);
+    state.early_delay_line.Initialize(delay_samples, work_buffer_ptr);
+
+    UpdateI3dl2Reverb(info, state, true);
+}
+
+void CommandGenerator::UpdateI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state,
+                                         bool should_clear) {
+
+    state.dry_gain = info.dry_gain;
+    state.shelf_filter.fill(0.0f);
+    state.lowpass_0 = 0.0f;
+    state.early_gain = Pow10(std::min(info.room + info.reflection, 5000.0f) / 2000.0f);
+    state.late_gain = Pow10(std::min(info.room + info.reverb, 5000.0f) / 2000.0f);
+
+    const auto sample_rate = info.sample_rate / 1000;
+    const f32 hf_gain = Pow10(info.room_hf / 2000.0f);
+    if (hf_gain >= 1.0f) {
+        state.lowpass_2 = 1.0f;
+        state.lowpass_1 = 0.0f;
+    } else {
+        const auto a = 1.0f - hf_gain;
+        const auto b = 2.0f * (1.0f - hf_gain * CosD(256.0f * info.hf_reference /
+                                                     static_cast<f32>(info.sample_rate)));
+        const auto c = std::sqrt(b * b - 4.0f * a * a);
+
+        state.lowpass_1 = (b - c) / (2.0f * a);
+        state.lowpass_2 = 1.0f - state.lowpass_1;
+    }
+    state.early_to_late_taps = AudioCommon::CalculateDelaySamples(
+        sample_rate, 1000.0f * (info.reflection_delay + info.reverb_delay));
+
+    state.last_reverb_echo = 0.6f * info.diffusion * 0.01f;
+    for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+        const auto length =
+            FDN_MIN_DELAY_LINE_TIMES[i] +
+            (info.density / 100.0f) * (FDN_MAX_DELAY_LINE_TIMES[i] - FDN_MIN_DELAY_LINE_TIMES[i]);
+        state.fdn_delay_line[i].SetDelay(AudioCommon::CalculateDelaySamples(sample_rate, length));
+
+        const auto delay_sample_counts = state.fdn_delay_line[i].GetDelay() +
+                                         state.decay_delay_line0[i].GetDelay() +
+                                         state.decay_delay_line1[i].GetDelay();
+
+        float a = (-60.0f * static_cast<f32>(delay_sample_counts)) /
+                  (info.decay_time * static_cast<f32>(info.sample_rate));
+        float b = a / info.hf_decay_ratio;
+        float c = CosD(128.0f * 0.5f * info.hf_reference / static_cast<f32>(info.sample_rate)) /
+                  SinD(128.0f * 0.5f * info.hf_reference / static_cast<f32>(info.sample_rate));
+        float d = Pow10((b - a) / 40.0f);
+        float e = Pow10((b + a) / 40.0f) * 0.7071f;
+
+        state.lpf_coefficients[0][i] = e * ((d * c) + 1.0f) / (c + d);
+        state.lpf_coefficients[1][i] = e * (1.0f - (d * c)) / (c + d);
+        state.lpf_coefficients[2][i] = (c - d) / (c + d);
+
+        state.decay_delay_line0[i].SetCoefficient(state.last_reverb_echo);
+        state.decay_delay_line1[i].SetCoefficient(-0.9f * state.last_reverb_echo);
+    }
+
+    if (should_clear) {
+        for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+            state.fdn_delay_line[i].Clear();
+            state.decay_delay_line0[i].Clear();
+            state.decay_delay_line1[i].Clear();
+        }
+        state.early_delay_line.Clear();
+        state.center_delay_line.Clear();
+    }
+
+    const auto max_early_delay = state.early_delay_line.GetMaxDelay();
+    const auto reflection_time = 1000.0f * (0.0098f * info.reverb_delay + 0.02f);
+    for (std::size_t tap = 0; tap < AudioCommon::I3DL2REVERB_TAPS; tap++) {
+        const auto length = AudioCommon::CalculateDelaySamples(
+            sample_rate, 1000.0f * info.reflection_delay + reflection_time * EARLY_TAP_TIMES[tap]);
+        state.early_tap_steps[tap] = std::min(length, max_early_delay);
+    }
+}
+
 void CommandGenerator::GenerateVolumeRampCommand(float last_volume, float current_volume,
                                                  s32 channel, s32 node_id) {
     const auto last = static_cast<s32>(last_volume * 32768.0f);

src/audio_core/command_generator.h

@@ -21,6 +21,8 @@ class ServerMixInfo;
 class EffectContext;
 class EffectBase;
 struct AuxInfoDSP;
+struct I3dl2ReverbParams;
+struct I3dl2ReverbState;
 using MixVolumeBuffer = std::array<float, AudioCommon::MAX_MIX_BUFFERS>;
 
 class CommandGenerator {
@@ -80,6 +82,9 @@ private:
     s32 ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u32 max_samples, s32* out_data,
                       u32 sample_count, u32 read_offset, u32 read_count);
 
+    void InitializeI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state,
+                               std::vector<u8>& work_buffer);
+    void UpdateI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state, bool should_clear);
     // DSP Code
     s32 DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 sample_count,
                     s32 channel, std::size_t mix_offset);

src/audio_core/common.h

@@ -33,6 +33,29 @@ constexpr std::size_t TEMP_MIX_BASE_SIZE = 0x3f00; // TODO(ogniK): Work out this
 // and our const ends up being 0x3f04, the 4 bytes are most
 // likely the sample history
 constexpr std::size_t TOTAL_TEMP_MIX_SIZE = TEMP_MIX_BASE_SIZE + AudioCommon::MAX_SAMPLE_HISTORY;
+constexpr f32 I3DL2REVERB_MAX_LEVEL = 5000.0f;
+constexpr f32 I3DL2REVERB_MIN_REFLECTION_DURATION = 0.02f;
+constexpr std::size_t I3DL2REVERB_TAPS = 20;
+constexpr std::size_t I3DL2REVERB_DELAY_LINE_COUNT = 4;
+using Fractional = s32;
+
+template <typename T>
+constexpr Fractional ToFractional(T x) {
+    return static_cast<Fractional>(x * static_cast<T>(0x4000));
+}
+
+constexpr Fractional MultiplyFractional(Fractional lhs, Fractional rhs) {
+    return static_cast<Fractional>(static_cast<s64>(lhs) * rhs >> 14);
+}
+
+constexpr s32 FractionalToFixed(Fractional x) {
+    const auto s = x & (1 << 13);
+    return static_cast<s32>(x >> 14) + s;
+}
+
+constexpr s32 CalculateDelaySamples(s32 sample_rate_khz, float time) {
+    return FractionalToFixed(MultiplyFractional(ToFractional(sample_rate_khz), ToFractional(time)));
+}
 
 static constexpr u32 VersionFromRevision(u32_le rev) {
     // "REV7" -> 7

src/audio_core/delay_line.cpp

@@ -0,0 +1,104 @@
+#include <cstring>
+#include "audio_core/delay_line.h"
+
+namespace AudioCore {
+DelayLineBase::DelayLineBase() = default;
+DelayLineBase::~DelayLineBase() = default;
+
+void DelayLineBase::Initialize(s32 max_delay_, float* src_buffer) {
+    buffer = src_buffer;
+    buffer_end = buffer + max_delay_;
+    max_delay = max_delay_;
+    output = buffer;
+    SetDelay(max_delay_);
+    Clear();
+}
+
+void DelayLineBase::SetDelay(s32 new_delay) {
+    if (max_delay < new_delay) {
+        return;
+    }
+    delay = new_delay;
+    input = (buffer + ((output - buffer) + new_delay) % (max_delay + 1));
+}
+
+s32 DelayLineBase::GetDelay() const {
+    return delay;
+}
+
+s32 DelayLineBase::GetMaxDelay() const {
+    return max_delay;
+}
+
+f32 DelayLineBase::TapOut(s32 last_sample) {
+    const float* ptr = input - (last_sample + 1);
+    if (ptr < buffer) {
+        ptr += (max_delay + 1);
+    }
+
+    return *ptr;
+}
+
+f32 DelayLineBase::Tick(f32 sample) {
+    *(input++) = sample;
+    const auto out_sample = *(output++);
+
+    if (buffer_end < input) {
+        input = buffer;
+    }
+
+    if (buffer_end < output) {
+        output = buffer;
+    }
+
+    return out_sample;
+}
+
+float* DelayLineBase::GetInput() {
+    return input;
+}
+
+const float* DelayLineBase::GetInput() const {
+    return input;
+}
+
+f32 DelayLineBase::GetOutputSample() const {
+    return *output;
+}
+
+void DelayLineBase::Clear() {
+    std::memset(buffer, 0, sizeof(float) * max_delay);
+}
+
+void DelayLineBase::Reset() {
+    buffer = nullptr;
+    buffer_end = nullptr;
+    max_delay = 0;
+    input = nullptr;
+    output = nullptr;
+    delay = 0;
+}
+
+DelayLineAllPass::DelayLineAllPass() = default;
+DelayLineAllPass::~DelayLineAllPass() = default;
+
+void DelayLineAllPass::Initialize(u32 delay_, float coeffcient_, f32* src_buffer) {
+    DelayLineBase::Initialize(delay_, src_buffer);
+    SetCoefficient(coeffcient_);
+}
+
+void DelayLineAllPass::SetCoefficient(float coeffcient_) {
+    coefficient = coeffcient_;
+}
+
+f32 DelayLineAllPass::Tick(f32 sample) {
+    const auto temp = sample - coefficient * *output;
+    return coefficient * temp + DelayLineBase::Tick(temp);
+}
+
+void DelayLineAllPass::Reset() {
+    coefficient = 0.0f;
+    DelayLineBase::Reset();
+}
+
+} // namespace AudioCore

src/audio_core/delay_line.h

@@ -0,0 +1,46 @@
+#pragma once
+
+#include "common/common_types.h"
+
+namespace AudioCore {
+
+class DelayLineBase {
+public:
+    DelayLineBase();
+    ~DelayLineBase();
+
+    void Initialize(s32 max_delay_, float* src_buffer);
+    void SetDelay(s32 new_delay);
+    s32 GetDelay() const;
+    s32 GetMaxDelay() const;
+    f32 TapOut(s32 last_sample);
+    f32 Tick(f32 sample);
+    float* GetInput();
+    const float* GetInput() const;
+    f32 GetOutputSample() const;
+    void Clear();
+    void Reset();
+
+protected:
+    float* buffer{nullptr};
+    float* buffer_end{nullptr};
+    s32 max_delay{};
+    float* input{nullptr};
+    float* output{nullptr};
+    s32 delay{};
+};
+
+class DelayLineAllPass final : public DelayLineBase {
+public:
+    DelayLineAllPass();
+    ~DelayLineAllPass();
+
+    void Initialize(u32 delay, float coeffcient_, f32* src_buffer);
+    void SetCoefficient(float coeffcient_);
+    f32 Tick(f32 sample);
+    void Reset();
+
+private:
+    float coefficient{};
+};
+} // namespace AudioCore

src/audio_core/effect_context.cpp

@@ -90,6 +90,14 @@ s32 EffectBase::GetProcessingOrder() const {
     return processing_order;
 }
 
+std::vector<u8>& EffectBase::GetWorkBuffer() {
+    return work_buffer;
+}
+
+const std::vector<u8>& EffectBase::GetWorkBuffer() const {
+    return work_buffer;
+}
+
 EffectI3dl2Reverb::EffectI3dl2Reverb() : EffectGeneric(EffectType::I3dl2Reverb) {}
 EffectI3dl2Reverb::~EffectI3dl2Reverb() = default;
 
@@ -117,6 +125,12 @@ void EffectI3dl2Reverb::Update(EffectInfo::InParams& in_params) {
         usage = UsageState::Initialized;
         params.status = ParameterStatus::Initialized;
         skipped = in_params.buffer_address == 0 || in_params.buffer_size == 0;
+        if (!skipped) {
+            auto& cur_work_buffer = GetWorkBuffer();
+            // Has two buffers internally
+            cur_work_buffer.resize(in_params.buffer_size * 2);
+            std::fill(cur_work_buffer.begin(), cur_work_buffer.end(), 0);
+        }
     }
 }
 
@@ -129,6 +143,14 @@ void EffectI3dl2Reverb::UpdateForCommandGeneration() {
     GetParams().status = ParameterStatus::Updated;
 }
 
+I3dl2ReverbState& EffectI3dl2Reverb::GetState() {
+    return state;
+}
+
+const I3dl2ReverbState& EffectI3dl2Reverb::GetState() const {
+    return state;
+}
+
 EffectBiquadFilter::EffectBiquadFilter() : EffectGeneric(EffectType::BiquadFilter) {}
 EffectBiquadFilter::~EffectBiquadFilter() = default;
 

src/audio_core/effect_context.h

@@ -8,6 +8,7 @@
 #include <memory>
 #include <vector>
 #include "audio_core/common.h"
+#include "audio_core/delay_line.h"
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/swap.h"
@@ -194,6 +195,8 @@ public:
     [[nodiscard]] bool IsEnabled() const;
     [[nodiscard]] s32 GetMixID() const;
     [[nodiscard]] s32 GetProcessingOrder() const;
+    [[nodiscard]] std::vector<u8>& GetWorkBuffer();
+    [[nodiscard]] const std::vector<u8>& GetWorkBuffer() const;
 
 protected:
     UsageState usage{UsageState::Invalid};
@@ -201,6 +204,7 @@ protected:
     s32 mix_id{};
     s32 processing_order{};
     bool enabled = false;
+    std::vector<u8> work_buffer{};
 };
 
 template <typename T>
@@ -212,7 +216,7 @@ public:
         return internal_params;
     }
 
-    const I3dl2ReverbParams& GetParams() const {
+    const T& GetParams() const {
         return internal_params;
     }
 
@@ -229,6 +233,27 @@ public:
     void UpdateForCommandGeneration() override;
 };
 
+struct I3dl2ReverbState {
+    f32 lowpass_0{};
+    f32 lowpass_1{};
+    f32 lowpass_2{};
+
+    DelayLineBase early_delay_line{};
+    std::array<u32, AudioCommon::I3DL2REVERB_TAPS> early_tap_steps{};
+    f32 early_gain{};
+    f32 late_gain{};
+
+    u32 early_to_late_taps{};
+    std::array<DelayLineBase, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> fdn_delay_line{};
+    std::array<DelayLineAllPass, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> decay_delay_line0{};
+    std::array<DelayLineAllPass, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> decay_delay_line1{};
+    f32 last_reverb_echo{};
+    DelayLineBase center_delay_line{};
+    std::array<std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT>, 3> lpf_coefficients{};
+    std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> shelf_filter{};
+    f32 dry_gain{};
+};
+
 class EffectI3dl2Reverb : public EffectGeneric<I3dl2ReverbParams> {
 public:
     explicit EffectI3dl2Reverb();
@@ -237,8 +262,12 @@ public:
     void Update(EffectInfo::InParams& in_params) override;
     void UpdateForCommandGeneration() override;
 
+    I3dl2ReverbState& GetState();
+    const I3dl2ReverbState& GetState() const;
+
 private:
     bool skipped = false;
+    I3dl2ReverbState state{};
 };
 
 class EffectBiquadFilter : public EffectGeneric<BiquadFilterParams> {

src/common/CMakeLists.txt

@@ -97,6 +97,7 @@ add_custom_command(OUTPUT scm_rev.cpp
 add_library(common STATIC
     algorithm.h
     alignment.h
+    assert.cpp
     assert.h
     atomic_ops.h
     detached_tasks.cpp
@@ -109,6 +110,7 @@ add_library(common STATIC
     cityhash.h
     common_funcs.h
     common_paths.h
+    common_sizes.h
     common_types.h
     concepts.h
     div_ceil.h
@@ -167,8 +169,8 @@ add_library(common STATIC
     threadsafe_queue.h
     time_zone.cpp
     time_zone.h
+    tiny_mt.h
     tree.h
-    uint128.cpp
     uint128.h
     uuid.cpp
     uuid.h

src/common/alignment.h

@@ -42,6 +42,11 @@ requires std::is_integral_v<T>[[nodiscard]] constexpr bool IsAligned(T value, si
     return (value & mask) == 0;
 }
 
+template <typename T, typename U>
+requires std::is_integral_v<T>[[nodiscard]] constexpr T DivideUp(T x, U y) {
+    return (x + (y - 1)) / y;
+}
+
 template <typename T, size_t Align = 16>
 class AlignmentAllocator {
 public:

src/common/assert.cpp

@@ -0,0 +1,11 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/assert.h"
+
+#include "common/common_funcs.h"
+
+void assert_handle_failure() {
+    Crash();
+}

src/common/assert.h

@@ -4,10 +4,13 @@
 
 #pragma once
 
-#include <cstdlib>
-#include "common/common_funcs.h"
 #include "common/logging/log.h"
 
+// Sometimes we want to try to continue even after hitting an assert.
+// However touching this file yields a global recompilation as this header is included almost
+// everywhere. So let's just move the handling of the failed assert to a single cpp file.
+void assert_handle_failure();
+
 // For asserts we'd like to keep all the junk executed when an assert happens away from the
 // important code in the function. One way of doing this is to put all the relevant code inside a
 // lambda and force the compiler to not inline it. Unfortunately, MSVC seems to have no syntax to
@@ -17,15 +20,14 @@
 // enough for our purposes.
 template <typename Fn>
 #if defined(_MSC_VER)
-[[msvc::noinline, noreturn]]
+[[msvc::noinline]]
 #elif defined(__GNUC__)
-[[gnu::cold, gnu::noinline, noreturn]]
+[[gnu::cold, gnu::noinline]]
 #endif
 static void
 assert_noinline_call(const Fn& fn) {
     fn();
-    Crash();
-    exit(1); // Keeps GCC's mouth shut about this actually returning
+    assert_handle_failure();
 }
 
 #define ASSERT(_a_)                                                                                \

src/common/cityhash.cpp

@@ -28,8 +28,10 @@
 // compromising on hash quality.
 
 #include <algorithm>
-#include <string.h> // for memcpy and memset
-#include "cityhash.h"
+#include <cstring>
+#include <utility>
+
+#include "common/cityhash.h"
 #include "common/swap.h"
 
 // #include "config.h"
@@ -42,21 +44,17 @@
 
 using namespace std;
 
-typedef uint8_t uint8;
-typedef uint32_t uint32;
-typedef uint64_t uint64;
-
 namespace Common {
 
-static uint64 UNALIGNED_LOAD64(const char* p) {
-    uint64 result;
-    memcpy(&result, p, sizeof(result));
+static u64 unaligned_load64(const char* p) {
+    u64 result;
+    std::memcpy(&result, p, sizeof(result));
     return result;
 }
 
-static uint32 UNALIGNED_LOAD32(const char* p) {
-    uint32 result;
-    memcpy(&result, p, sizeof(result));
+static u32 unaligned_load32(const char* p) {
+    u32 result;
+    std::memcpy(&result, p, sizeof(result));
     return result;
 }
 
@@ -76,64 +74,64 @@ static uint32 UNALIGNED_LOAD32(const char* p) {
 #endif
 #endif
 
-static uint64 Fetch64(const char* p) {
-    return uint64_in_expected_order(UNALIGNED_LOAD64(p));
+static u64 Fetch64(const char* p) {
+    return uint64_in_expected_order(unaligned_load64(p));
 }
 
-static uint32 Fetch32(const char* p) {
-    return uint32_in_expected_order(UNALIGNED_LOAD32(p));
+static u32 Fetch32(const char* p) {
+    return uint32_in_expected_order(unaligned_load32(p));
 }
 
 // Some primes between 2^63 and 2^64 for various uses.
-static const uint64 k0 = 0xc3a5c85c97cb3127ULL;
-static const uint64 k1 = 0xb492b66fbe98f273ULL;
-static const uint64 k2 = 0x9ae16a3b2f90404fULL;
+static constexpr u64 k0 = 0xc3a5c85c97cb3127ULL;
+static constexpr u64 k1 = 0xb492b66fbe98f273ULL;
+static constexpr u64 k2 = 0x9ae16a3b2f90404fULL;
 
 // Bitwise right rotate.  Normally this will compile to a single
 // instruction, especially if the shift is a manifest constant.
-static uint64 Rotate(uint64 val, int shift) {
+static u64 Rotate(u64 val, int shift) {
     // Avoid shifting by 64: doing so yields an undefined result.
     return shift == 0 ? val : ((val >> shift) | (val << (64 - shift)));
 }
 
-static uint64 ShiftMix(uint64 val) {
+static u64 ShiftMix(u64 val) {
     return val ^ (val >> 47);
 }
 
-static uint64 HashLen16(uint64 u, uint64 v) {
-    return Hash128to64(uint128(u, v));
+static u64 HashLen16(u64 u, u64 v) {
+    return Hash128to64(u128{u, v});
 }
 
-static uint64 HashLen16(uint64 u, uint64 v, uint64 mul) {
+static u64 HashLen16(u64 u, u64 v, u64 mul) {
     // Murmur-inspired hashing.
-    uint64 a = (u ^ v) * mul;
+    u64 a = (u ^ v) * mul;
     a ^= (a >> 47);
-    uint64 b = (v ^ a) * mul;
+    u64 b = (v ^ a) * mul;
     b ^= (b >> 47);
     b *= mul;
     return b;
 }
 
-static uint64 HashLen0to16(const char* s, std::size_t len) {
+static u64 HashLen0to16(const char* s, size_t len) {
     if (len >= 8) {
-        uint64 mul = k2 + len * 2;
-        uint64 a = Fetch64(s) + k2;
-        uint64 b = Fetch64(s + len - 8);
-        uint64 c = Rotate(b, 37) * mul + a;
-        uint64 d = (Rotate(a, 25) + b) * mul;
+        u64 mul = k2 + len * 2;
+        u64 a = Fetch64(s) + k2;
+        u64 b = Fetch64(s + len - 8);
+        u64 c = Rotate(b, 37) * mul + a;
+        u64 d = (Rotate(a, 25) + b) * mul;
         return HashLen16(c, d, mul);
     }
     if (len >= 4) {
-        uint64 mul = k2 + len * 2;
-        uint64 a = Fetch32(s);
+        u64 mul = k2 + len * 2;
+        u64 a = Fetch32(s);
         return HashLen16(len + (a << 3), Fetch32(s + len - 4), mul);
     }
     if (len > 0) {
-        uint8 a = s[0];
-        uint8 b = s[len >> 1];
-        uint8 c = s[len - 1];
-        uint32 y = static_cast<uint32>(a) + (static_cast<uint32>(b) << 8);
-        uint32 z = static_cast<uint32>(len) + (static_cast<uint32>(c) << 2);
+        u8 a = s[0];
+        u8 b = s[len >> 1];
+        u8 c = s[len - 1];
+        u32 y = static_cast<u32>(a) + (static_cast<u32>(b) << 8);
+        u32 z = static_cast<u32>(len) + (static_cast<u32>(c) << 2);
         return ShiftMix(y * k2 ^ z * k0) * k2;
     }
     return k2;
@@ -141,22 +139,21 @@ static uint64 HashLen0to16(const char* s, std::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, std::size_t len) {
-    uint64 mul = k2 + len * 2;
-    uint64 a = Fetch64(s) * k1;
-    uint64 b = Fetch64(s + 8);
-    uint64 c = Fetch64(s + len - 8) * mul;
-    uint64 d = Fetch64(s + len - 16) * k2;
+static u64 HashLen17to32(const char* s, size_t len) {
+    u64 mul = k2 + len * 2;
+    u64 a = Fetch64(s) * k1;
+    u64 b = Fetch64(s + 8);
+    u64 c = Fetch64(s + len - 8) * mul;
+    u64 d = Fetch64(s + len - 16) * k2;
     return HashLen16(Rotate(a + b, 43) + Rotate(c, 30) + d, a + Rotate(b + k2, 18) + c, mul);
 }
 
 // Return a 16-byte hash for 48 bytes.  Quick and dirty.
 // Callers do best to use "random-looking" values for a and b.
-static pair<uint64, uint64> WeakHashLen32WithSeeds(uint64 w, uint64 x, uint64 y, uint64 z, uint64 a,
-                                                   uint64 b) {
+static pair<u64, u64> WeakHashLen32WithSeeds(u64 w, u64 x, u64 y, u64 z, u64 a, u64 b) {
     a += w;
     b = Rotate(b + a + z, 21);
-    uint64 c = a;
+    u64 c = a;
     a += x;
     a += y;
     b += Rotate(a, 44);
@@ -164,34 +161,34 @@ static pair<uint64, uint64> WeakHashLen32WithSeeds(uint64 w, uint64 x, uint64 y,
 }
 
 // Return a 16-byte hash for s[0] ... s[31], a, and b.  Quick and dirty.
-static pair<uint64, uint64> WeakHashLen32WithSeeds(const char* s, uint64 a, uint64 b) {
+static pair<u64, u64> WeakHashLen32WithSeeds(const char* s, u64 a, u64 b) {
     return WeakHashLen32WithSeeds(Fetch64(s), Fetch64(s + 8), Fetch64(s + 16), Fetch64(s + 24), a,
                                   b);
 }
 
 // Return an 8-byte hash for 33 to 64 bytes.
-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);
-    uint64 c = Fetch64(s + len - 24);
-    uint64 d = Fetch64(s + len - 32);
-    uint64 e = Fetch64(s + 16) * k2;
-    uint64 f = Fetch64(s + 24) * 9;
-    uint64 g = Fetch64(s + len - 8);
-    uint64 h = Fetch64(s + len - 16) * mul;
-    uint64 u = Rotate(a + g, 43) + (Rotate(b, 30) + c) * 9;
-    uint64 v = ((a + g) ^ d) + f + 1;
-    uint64 w = swap64((u + v) * mul) + h;
-    uint64 x = Rotate(e + f, 42) + c;
-    uint64 y = (swap64((v + w) * mul) + g) * mul;
-    uint64 z = e + f + c;
+static u64 HashLen33to64(const char* s, size_t len) {
+    u64 mul = k2 + len * 2;
+    u64 a = Fetch64(s) * k2;
+    u64 b = Fetch64(s + 8);
+    u64 c = Fetch64(s + len - 24);
+    u64 d = Fetch64(s + len - 32);
+    u64 e = Fetch64(s + 16) * k2;
+    u64 f = Fetch64(s + 24) * 9;
+    u64 g = Fetch64(s + len - 8);
+    u64 h = Fetch64(s + len - 16) * mul;
+    u64 u = Rotate(a + g, 43) + (Rotate(b, 30) + c) * 9;
+    u64 v = ((a + g) ^ d) + f + 1;
+    u64 w = swap64((u + v) * mul) + h;
+    u64 x = Rotate(e + f, 42) + c;
+    u64 y = (swap64((v + w) * mul) + g) * mul;
+    u64 z = e + f + c;
     a = swap64((x + z) * mul + y) + b;
     b = ShiftMix((z + a) * mul + d + h) * mul;
     return b + x;
 }
 
-uint64 CityHash64(const char* s, std::size_t len) {
+u64 CityHash64(const char* s, size_t len) {
     if (len <= 32) {
         if (len <= 16) {
             return HashLen0to16(s, len);
@@ -204,15 +201,15 @@ uint64 CityHash64(const char* s, std::size_t len) {
 
     // For strings over 64 bytes we hash the end first, and then as we
     // loop we keep 56 bytes of state: v, w, x, y, and z.
-    uint64 x = Fetch64(s + len - 40);
-    uint64 y = Fetch64(s + len - 16) + Fetch64(s + len - 56);
-    uint64 z = HashLen16(Fetch64(s + len - 48) + len, Fetch64(s + len - 24));
-    pair<uint64, uint64> v = WeakHashLen32WithSeeds(s + len - 64, len, z);
-    pair<uint64, uint64> w = WeakHashLen32WithSeeds(s + len - 32, y + k1, x);
+    u64 x = Fetch64(s + len - 40);
+    u64 y = Fetch64(s + len - 16) + Fetch64(s + len - 56);
+    u64 z = HashLen16(Fetch64(s + len - 48) + len, Fetch64(s + len - 24));
+    pair<u64, u64> v = WeakHashLen32WithSeeds(s + len - 64, len, z);
+    pair<u64, u64> w = WeakHashLen32WithSeeds(s + len - 32, y + k1, x);
     x = x * k1 + Fetch64(s);
 
     // Decrease len to the nearest multiple of 64, and operate on 64-byte chunks.
-    len = (len - 1) & ~static_cast<std::size_t>(63);
+    len = (len - 1) & ~static_cast<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;
@@ -229,21 +226,21 @@ uint64 CityHash64(const char* s, std::size_t len) {
                      HashLen16(v.second, w.second) + x);
 }
 
-uint64 CityHash64WithSeed(const char* s, std::size_t len, uint64 seed) {
+u64 CityHash64WithSeed(const char* s, size_t len, u64 seed) {
     return CityHash64WithSeeds(s, len, k2, seed);
 }
 
-uint64 CityHash64WithSeeds(const char* s, std::size_t len, uint64 seed0, uint64 seed1) {
+u64 CityHash64WithSeeds(const char* s, size_t len, u64 seed0, u64 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, std::size_t len, uint128 seed) {
-    uint64 a = Uint128Low64(seed);
-    uint64 b = Uint128High64(seed);
-    uint64 c = 0;
-    uint64 d = 0;
+static u128 CityMurmur(const char* s, size_t len, u128 seed) {
+    u64 a = seed[0];
+    u64 b = seed[1];
+    u64 c = 0;
+    u64 d = 0;
     signed long l = static_cast<long>(len) - 16;
     if (l <= 0) { // len <= 16
         a = ShiftMix(a * k1) * k1;
@@ -266,20 +263,20 @@ static uint128 CityMurmur(const char* s, std::size_t len, uint128 seed) {
     }
     a = HashLen16(a, c);
     b = HashLen16(d, b);
-    return uint128(a ^ b, HashLen16(b, a));
+    return u128{a ^ b, HashLen16(b, a)};
 }
 
-uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed) {
+u128 CityHash128WithSeed(const char* s, size_t len, u128 seed) {
     if (len < 128) {
         return CityMurmur(s, len, seed);
     }
 
     // We expect len >= 128 to be the common case.  Keep 56 bytes of state:
     // v, w, x, y, and z.
-    pair<uint64, uint64> v, w;
-    uint64 x = Uint128Low64(seed);
-    uint64 y = Uint128High64(seed);
-    uint64 z = len * k1;
+    pair<u64, u64> v, w;
+    u64 x = seed[0];
+    u64 y = seed[1];
+    u64 z = len * k1;
     v.first = Rotate(y ^ k1, 49) * k1 + Fetch64(s);
     v.second = Rotate(v.first, 42) * k1 + Fetch64(s + 8);
     w.first = Rotate(y + z, 35) * k1 + x;
@@ -313,7 +310,7 @@ uint128 CityHash128WithSeed(const char* s, std::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 (std::size_t tail_done = 0; tail_done < len;) {
+    for (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);
@@ -328,13 +325,12 @@ uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed) {
     // different 56-byte-to-8-byte hashes to get a 16-byte final result.
     x = HashLen16(x, v.first);
     y = HashLen16(y + z, w.first);
-    return uint128(HashLen16(x + v.second, w.second) + y, HashLen16(x + w.second, y + v.second));
+    return u128{HashLen16(x + v.second, w.second) + y, HashLen16(x + w.second, y + v.second)};
 }
 
-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));
+u128 CityHash128(const char* s, size_t len) {
+    return len >= 16 ? CityHash128WithSeed(s + 16, len - 16, u128{Fetch64(s), Fetch64(s + 8) + k0})
+                     : CityHash128WithSeed(s, len, u128{k0, k1});
 }
 
 } // namespace Common

src/common/cityhash.h

@@ -62,49 +62,38 @@
 #pragma once
 
 #include <cstddef>
-#include <cstdint>
-#include <utility>
+#include "common/common_types.h"
 
 namespace Common {
 
-using uint128 = std::pair<uint64_t, uint64_t>;
-
-[[nodiscard]] inline uint64_t Uint128Low64(const uint128& x) {
-    return x.first;
-}
-[[nodiscard]] inline uint64_t Uint128High64(const uint128& x) {
-    return x.second;
-}
-
 // Hash function for a byte array.
-[[nodiscard]] uint64_t CityHash64(const char* buf, std::size_t len);
+[[nodiscard]] u64 CityHash64(const char* buf, size_t len);
 
 // Hash function for a byte array.  For convenience, a 64-bit seed is also
 // hashed into the result.
-[[nodiscard]] uint64_t CityHash64WithSeed(const char* buf, std::size_t len, uint64_t seed);
+[[nodiscard]] u64 CityHash64WithSeed(const char* buf, size_t len, u64 seed);
 
 // Hash function for a byte array.  For convenience, two seeds are also
 // hashed into the result.
-[[nodiscard]] uint64_t CityHash64WithSeeds(const char* buf, std::size_t len, uint64_t seed0,
-                                           uint64_t seed1);
+[[nodiscard]] u64 CityHash64WithSeeds(const char* buf, size_t len, u64 seed0, u64 seed1);
 
 // Hash function for a byte array.
-[[nodiscard]] uint128 CityHash128(const char* s, std::size_t len);
+[[nodiscard]] u128 CityHash128(const char* s, size_t len);
 
 // Hash function for a byte array.  For convenience, a 128-bit seed is also
 // hashed into the result.
-[[nodiscard]] uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed);
+[[nodiscard]] u128 CityHash128WithSeed(const char* s, size_t len, u128 seed);
 
 // Hash 128 input bits down to 64 bits of output.
 // This is intended to be a reasonably good hash function.
-[[nodiscard]] inline uint64_t Hash128to64(const uint128& x) {
+[[nodiscard]] inline u64 Hash128to64(const u128& x) {
     // Murmur-inspired hashing.
-    const uint64_t kMul = 0x9ddfea08eb382d69ULL;
-    uint64_t a = (Uint128Low64(x) ^ Uint128High64(x)) * kMul;
+    const u64 mul = 0x9ddfea08eb382d69ULL;
+    u64 a = (x[0] ^ x[1]) * mul;
     a ^= (a >> 47);
-    uint64_t b = (Uint128High64(x) ^ a) * kMul;
+    u64 b = (x[1] ^ a) * mul;
     b ^= (b >> 47);
-    b *= kMul;
+    b *= mul;
     return b;
 }
 

src/common/common_funcs.h

@@ -52,9 +52,13 @@ __declspec(dllimport) void __stdcall DebugBreak(void);
 // Generic function to get last error message.
 // Call directly after the command or use the error num.
 // This function might change the error code.
-// Defined in Misc.cpp.
+// Defined in misc.cpp.
 [[nodiscard]] std::string GetLastErrorMsg();
 
+// Like GetLastErrorMsg(), but passing an explicit error code.
+// Defined in misc.cpp.
+[[nodiscard]] std::string NativeErrorToString(int e);
+
 #define DECLARE_ENUM_FLAG_OPERATORS(type)                                                          \
     [[nodiscard]] constexpr type operator|(type a, type b) noexcept {                              \
         using T = std::underlying_type_t<type>;                                                    \

src/common/common_sizes.h

@@ -0,0 +1,43 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <limits>
+
+#include "common/common_types.h"
+
+namespace Common {
+
+enum : u64 {
+    Size_1_KB = 0x400ULL,
+    Size_64_KB = 64ULL * Size_1_KB,
+    Size_128_KB = 128ULL * Size_1_KB,
+    Size_1_MB = 0x100000ULL,
+    Size_2_MB = 2ULL * Size_1_MB,
+    Size_4_MB = 4ULL * Size_1_MB,
+    Size_5_MB = 5ULL * Size_1_MB,
+    Size_14_MB = 14ULL * Size_1_MB,
+    Size_32_MB = 32ULL * Size_1_MB,
+    Size_33_MB = 33ULL * Size_1_MB,
+    Size_128_MB = 128ULL * Size_1_MB,
+    Size_448_MB = 448ULL * Size_1_MB,
+    Size_507_MB = 507ULL * Size_1_MB,
+    Size_562_MB = 562ULL * Size_1_MB,
+    Size_1554_MB = 1554ULL * Size_1_MB,
+    Size_2048_MB = 2048ULL * Size_1_MB,
+    Size_2193_MB = 2193ULL * Size_1_MB,
+    Size_3285_MB = 3285ULL * Size_1_MB,
+    Size_4916_MB = 4916ULL * Size_1_MB,
+    Size_1_GB = 0x40000000ULL,
+    Size_2_GB = 2ULL * Size_1_GB,
+    Size_4_GB = 4ULL * Size_1_GB,
+    Size_6_GB = 6ULL * Size_1_GB,
+    Size_8_GB = 8ULL * Size_1_GB,
+    Size_64_GB = 64ULL * Size_1_GB,
+    Size_512_GB = 512ULL * Size_1_GB,
+    Size_Invalid = std::numeric_limits<u64>::max(),
+};
+
+} // namespace Common

src/common/fiber.cpp

@@ -11,7 +11,7 @@
 
 namespace Common {
 
-constexpr std::size_t default_stack_size = 256 * 1024;
+constexpr std::size_t default_stack_size = 512 * 1024;
 
 struct Fiber::FiberImpl {
     FiberImpl() : stack{default_stack_size}, rewind_stack{default_stack_size} {}
@@ -116,16 +116,19 @@ void Fiber::Rewind() {
     boost::context::detail::jump_fcontext(impl->rewind_context, this);
 }
 
-void Fiber::YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to) {
-    ASSERT_MSG(from != nullptr, "Yielding fiber is null!");
-    ASSERT_MSG(to != nullptr, "Next fiber is null!");
-    to->impl->guard.lock();
-    to->impl->previous_fiber = from;
-    auto transfer = boost::context::detail::jump_fcontext(to->impl->context, to.get());
-    ASSERT(from->impl->previous_fiber != nullptr);
-    from->impl->previous_fiber->impl->context = transfer.fctx;
-    from->impl->previous_fiber->impl->guard.unlock();
-    from->impl->previous_fiber.reset();
+void Fiber::YieldTo(std::weak_ptr<Fiber> weak_from, Fiber& to) {
+    to.impl->guard.lock();
+    to.impl->previous_fiber = weak_from.lock();
+
+    auto transfer = boost::context::detail::jump_fcontext(to.impl->context, &to);
+
+    // "from" might no longer be valid if the thread was killed
+    if (auto from = weak_from.lock()) {
+        ASSERT(from->impl->previous_fiber != nullptr);
+        from->impl->previous_fiber->impl->context = transfer.fctx;
+        from->impl->previous_fiber->impl->guard.unlock();
+        from->impl->previous_fiber.reset();
+    }
 }
 
 std::shared_ptr<Fiber> Fiber::ThreadToFiber() {

src/common/fiber.h

@@ -41,7 +41,7 @@ public:
 
     /// Yields control from Fiber 'from' to Fiber 'to'
     /// Fiber 'from' must be the currently running fiber.
-    static void YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to);
+    static void YieldTo(std::weak_ptr<Fiber> weak_from, Fiber& to);
     [[nodiscard]] static std::shared_ptr<Fiber> ThreadToFiber();
 
     void SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param);

src/common/logging/backend.cpp

@@ -212,6 +212,7 @@ void DebuggerBackend::Write(const Entry& entry) {
     SUB(Service, ARP)                                                                              \
     SUB(Service, BCAT)                                                                             \
     SUB(Service, BPC)                                                                              \
+    SUB(Service, BGTC)                                                                             \
     SUB(Service, BTDRV)                                                                            \
     SUB(Service, BTM)                                                                              \
     SUB(Service, Capture)                                                                          \

src/common/logging/log.h

@@ -66,6 +66,7 @@ enum class Class : ClassType {
     Service_ARP,       ///< The ARP service
     Service_Audio,     ///< The Audio (Audio control) service
     Service_BCAT,      ///< The BCAT service
+    Service_BGTC,      ///< The BGTC (Background Task Controller) service
     Service_BPC,       ///< The BPC service
     Service_BTDRV,     ///< The Bluetooth driver service
     Service_BTM,       ///< The BTM service

src/common/misc.cpp

@@ -12,27 +12,41 @@
 
 #include "common/common_funcs.h"
 
-// Generic function to get last error message.
-// Call directly after the command or use the error num.
-// This function might change the error code.
-std::string GetLastErrorMsg() {
-    static constexpr std::size_t buff_size = 255;
-    char err_str[buff_size];
-
+std::string NativeErrorToString(int e) {
 #ifdef _WIN32
-    FormatMessageA(FORMAT_MESSAGE_FROM_SYSTEM, nullptr, GetLastError(),
-                   MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), err_str, buff_size, nullptr);
-    return std::string(err_str, buff_size);
-#elif defined(__GLIBC__) && (_GNU_SOURCE || (_POSIX_C_SOURCE < 200112L && _XOPEN_SOURCE < 600))
+    LPSTR err_str;
+
+    DWORD res = FormatMessageA(FORMAT_MESSAGE_FROM_SYSTEM | FORMAT_MESSAGE_ALLOCATE_BUFFER |
+                                   FORMAT_MESSAGE_IGNORE_INSERTS,
+                               nullptr, e, MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT),
+                               reinterpret_cast<LPSTR>(&err_str), 1, nullptr);
+    if (!res) {
+        return "(FormatMessageA failed to format error)";
+    }
+    std::string ret(err_str);
+    LocalFree(err_str);
+    return ret;
+#else
+    char err_str[255];
+#if defined(__GLIBC__) && (_GNU_SOURCE || (_POSIX_C_SOURCE < 200112L && _XOPEN_SOURCE < 600))
     // Thread safe (GNU-specific)
-    const char* str = strerror_r(errno, err_str, buff_size);
+    const char* str = strerror_r(e, err_str, sizeof(err_str));
     return std::string(str);
 #else
     // Thread safe (XSI-compliant)
-    const int success = strerror_r(errno, err_str, buff_size);
-    if (success != 0) {
-        return {};
+    int second_err = strerror_r(e, err_str, sizeof(err_str));
+    if (second_err != 0) {
+        return "(strerror_r failed to format error)";
     }
     return std::string(err_str);
+#endif // GLIBC etc.
+#endif // _WIN32
+}
+
+std::string GetLastErrorMsg() {
+#ifdef _WIN32
+    return NativeErrorToString(GetLastError());
+#else
+    return NativeErrorToString(errno);
 #endif
 }

src/common/nvidia_flags.h

@@ -2,6 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#pragma once
+
 namespace Common {
 
 /// Configure platform specific flags for Nvidia's driver

src/common/threadsafe_queue.h

@@ -83,11 +83,15 @@ public:
         return true;
     }
 
-    T PopWait() {
+    void Wait() {
         if (Empty()) {
             std::unique_lock lock{cv_mutex};
             cv.wait(lock, [this]() { return !Empty(); });
         }
+    }
+
+    T PopWait() {
+        Wait();
         T t;
         Pop(t);
         return t;
@@ -156,6 +160,10 @@ public:
         return spsc_queue.Pop(t);
     }
 
+    void Wait() {
+        spsc_queue.Wait();
+    }
+
     T PopWait() {
         return spsc_queue.PopWait();
     }

src/common/tiny_mt.h

@@ -0,0 +1,250 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+
+#include "common/alignment.h"
+#include "common/common_types.h"
+
+namespace Common {
+
+// Implementation of TinyMT (mersenne twister RNG).
+// Like Nintendo, we will use the sample parameters.
+class TinyMT {
+public:
+    static constexpr std::size_t NumStateWords = 4;
+
+    struct State {
+        std::array<u32, NumStateWords> data{};
+    };
+
+private:
+    static constexpr u32 ParamMat1 = 0x8F7011EE;
+    static constexpr u32 ParamMat2 = 0xFC78FF1F;
+    static constexpr u32 ParamTmat = 0x3793FDFF;
+
+    static constexpr u32 ParamMult = 0x6C078965;
+    static constexpr u32 ParamPlus = 0x0019660D;
+    static constexpr u32 ParamXor = 0x5D588B65;
+
+    static constexpr u32 TopBitmask = 0x7FFFFFFF;
+
+    static constexpr int MinimumInitIterations = 8;
+    static constexpr int NumDiscardedInitOutputs = 8;
+
+    static constexpr u32 XorByShifted27(u32 value) {
+        return value ^ (value >> 27);
+    }
+
+    static constexpr u32 XorByShifted30(u32 value) {
+        return value ^ (value >> 30);
+    }
+
+private:
+    State state{};
+
+private:
+    // Internal API.
+    void FinalizeInitialization() {
+        const u32 state0 = this->state.data[0] & TopBitmask;
+        const u32 state1 = this->state.data[1];
+        const u32 state2 = this->state.data[2];
+        const u32 state3 = this->state.data[3];
+
+        if (state0 == 0 && state1 == 0 && state2 == 0 && state3 == 0) {
+            this->state.data[0] = 'T';
+            this->state.data[1] = 'I';
+            this->state.data[2] = 'N';
+            this->state.data[3] = 'Y';
+        }
+
+        for (int i = 0; i < NumDiscardedInitOutputs; i++) {
+            this->GenerateRandomU32();
+        }
+    }
+
+    u32 GenerateRandomU24() {
+        return (this->GenerateRandomU32() >> 8);
+    }
+
+    static void GenerateInitialValuePlus(TinyMT::State* state, int index, u32 value) {
+        u32& state0 = state->data[(index + 0) % NumStateWords];
+        u32& state1 = state->data[(index + 1) % NumStateWords];
+        u32& state2 = state->data[(index + 2) % NumStateWords];
+        u32& state3 = state->data[(index + 3) % NumStateWords];
+
+        const u32 x = XorByShifted27(state0 ^ state1 ^ state3) * ParamPlus;
+        const u32 y = x + index + value;
+
+        state0 = y;
+        state1 += x;
+        state2 += y;
+    }
+
+    static void GenerateInitialValueXor(TinyMT::State* state, int index) {
+        u32& state0 = state->data[(index + 0) % NumStateWords];
+        u32& state1 = state->data[(index + 1) % NumStateWords];
+        u32& state2 = state->data[(index + 2) % NumStateWords];
+        u32& state3 = state->data[(index + 3) % NumStateWords];
+
+        const u32 x = XorByShifted27(state0 + state1 + state3) * ParamXor;
+        const u32 y = x - index;
+
+        state0 = y;
+        state1 ^= x;
+        state2 ^= y;
+    }
+
+public:
+    constexpr TinyMT() = default;
+
+    // Public API.
+
+    // Initialization.
+    void Initialize(u32 seed) {
+        this->state.data[0] = seed;
+        this->state.data[1] = ParamMat1;
+        this->state.data[2] = ParamMat2;
+        this->state.data[3] = ParamTmat;
+
+        for (int i = 1; i < MinimumInitIterations; i++) {
+            const u32 mixed = XorByShifted30(this->state.data[(i - 1) % NumStateWords]);
+            this->state.data[i % NumStateWords] ^= mixed * ParamMult + i;
+        }
+
+        this->FinalizeInitialization();
+    }
+
+    void Initialize(const u32* seed, int seed_count) {
+        this->state.data[0] = 0;
+        this->state.data[1] = ParamMat1;
+        this->state.data[2] = ParamMat2;
+        this->state.data[3] = ParamTmat;
+
+        {
+            const int num_init_iterations = std::max(seed_count + 1, MinimumInitIterations) - 1;
+
+            GenerateInitialValuePlus(&this->state, 0, seed_count);
+
+            for (int i = 0; i < num_init_iterations; i++) {
+                GenerateInitialValuePlus(&this->state, (i + 1) % NumStateWords,
+                                         (i < seed_count) ? seed[i] : 0);
+            }
+
+            for (int i = 0; i < static_cast<int>(NumStateWords); i++) {
+                GenerateInitialValueXor(&this->state,
+                                        (i + 1 + num_init_iterations) % NumStateWords);
+            }
+        }
+
+        this->FinalizeInitialization();
+    }
+
+    // State management.
+    void GetState(TinyMT::State& out) const {
+        out.data = this->state.data;
+    }
+
+    void SetState(const TinyMT::State& state_) {
+        this->state.data = state_.data;
+    }
+
+    // Random generation.
+    void GenerateRandomBytes(void* dst, std::size_t size) {
+        const uintptr_t start = reinterpret_cast<uintptr_t>(dst);
+        const uintptr_t end = start + size;
+        const uintptr_t aligned_start = Common::AlignUp(start, 4);
+        const uintptr_t aligned_end = Common::AlignDown(end, 4);
+
+        // Make sure we're aligned.
+        if (start < aligned_start) {
+            const u32 rnd = this->GenerateRandomU32();
+            std::memcpy(dst, &rnd, aligned_start - start);
+        }
+
+        // Write as many aligned u32s as we can.
+        {
+            u32* cur_dst = reinterpret_cast<u32*>(aligned_start);
+            u32* const end_dst = reinterpret_cast<u32*>(aligned_end);
+
+            while (cur_dst < end_dst) {
+                *(cur_dst++) = this->GenerateRandomU32();
+            }
+        }
+
+        // Handle any leftover unaligned data.
+        if (aligned_end < end) {
+            const u32 rnd = this->GenerateRandomU32();
+            std::memcpy(reinterpret_cast<void*>(aligned_end), &rnd, end - aligned_end);
+        }
+    }
+
+    u32 GenerateRandomU32() {
+        // Advance state.
+        const u32 x0 =
+            (this->state.data[0] & TopBitmask) ^ this->state.data[1] ^ this->state.data[2];
+        const u32 y0 = this->state.data[3];
+        const u32 x1 = x0 ^ (x0 << 1);
+        const u32 y1 = y0 ^ (y0 >> 1) ^ x1;
+
+        const u32 state0 = this->state.data[1];
+        u32 state1 = this->state.data[2];
+        u32 state2 = x1 ^ (y1 << 10);
+        const u32 state3 = y1;
+
+        if ((y1 & 1) != 0) {
+            state1 ^= ParamMat1;
+            state2 ^= ParamMat2;
+        }
+
+        this->state.data[0] = state0;
+        this->state.data[1] = state1;
+        this->state.data[2] = state2;
+        this->state.data[3] = state3;
+
+        // Temper.
+        const u32 t1 = state0 + (state2 >> 8);
+        u32 t0 = state3 ^ t1;
+
+        if ((t1 & 1) != 0) {
+            t0 ^= ParamTmat;
+        }
+
+        return t0;
+    }
+
+    u64 GenerateRandomU64() {
+        const u32 lo = this->GenerateRandomU32();
+        const u32 hi = this->GenerateRandomU32();
+        return (u64{hi} << 32) | u64{lo};
+    }
+
+    float GenerateRandomF32() {
+        // Floats have 24 bits of mantissa.
+        constexpr u32 MantissaBits = 24;
+        return static_cast<float>(GenerateRandomU24()) * (1.0f / (1U << MantissaBits));
+    }
+
+    double GenerateRandomF64() {
+        // Doubles have 53 bits of mantissa.
+        // The smart way to generate 53 bits of random would be to use 32 bits
+        // from the first rnd32() call, and then 21 from the second.
+        // Nintendo does not. They use (32 - 5) = 27 bits from the first rnd32()
+        // call, and (32 - 6) bits from the second. We'll do what they do, but
+        // There's not a clear reason why.
+        constexpr u32 MantissaBits = 53;
+        constexpr u32 Shift1st = (64 - MantissaBits) / 2;
+        constexpr u32 Shift2nd = (64 - MantissaBits) - Shift1st;
+
+        const u32 first = (this->GenerateRandomU32() >> Shift1st);
+        const u32 second = (this->GenerateRandomU32() >> Shift2nd);
+
+        return (1.0 * first * (u64{1} << (32 - Shift2nd)) + second) *
+               (1.0 / (u64{1} << MantissaBits));
+    }
+};
+
+} // namespace Common

src/common/uint128.cpp

@@ -1,71 +0,0 @@
-// Copyright 2019 yuzu Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#ifdef _MSC_VER
-#include <intrin.h>
-
-#pragma intrinsic(_umul128)
-#pragma intrinsic(_udiv128)
-#endif
-#include <cstring>
-#include "common/uint128.h"
-
-namespace Common {
-
-#ifdef _MSC_VER
-
-u64 MultiplyAndDivide64(u64 a, u64 b, u64 d) {
-    u128 r{};
-    r[0] = _umul128(a, b, &r[1]);
-    u64 remainder;
-#if _MSC_VER < 1923
-    return udiv128(r[1], r[0], d, &remainder);
-#else
-    return _udiv128(r[1], r[0], d, &remainder);
-#endif
-}
-
-#else
-
-u64 MultiplyAndDivide64(u64 a, u64 b, u64 d) {
-    const u64 diva = a / d;
-    const u64 moda = a % d;
-    const u64 divb = b / d;
-    const u64 modb = b % d;
-    return diva * b + moda * divb + moda * modb / d;
-}
-
-#endif
-
-u128 Multiply64Into128(u64 a, u64 b) {
-    u128 result;
-#ifdef _MSC_VER
-    result[0] = _umul128(a, b, &result[1]);
-#else
-    unsigned __int128 tmp = a;
-    tmp *= b;
-    std::memcpy(&result, &tmp, sizeof(u128));
-#endif
-    return result;
-}
-
-std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor) {
-    u64 remainder = dividend[0] % divisor;
-    u64 accum = dividend[0] / divisor;
-    if (dividend[1] == 0)
-        return {accum, remainder};
-    // We ignore dividend[1] / divisor as that overflows
-    const u64 first_segment = (dividend[1] % divisor) << 32;
-    accum += (first_segment / divisor) << 32;
-    const u64 second_segment = (first_segment % divisor) << 32;
-    accum += (second_segment / divisor);
-    remainder += second_segment % divisor;
-    if (remainder >= divisor) {
-        accum++;
-        remainder -= divisor;
-    }
-    return {accum, remainder};
-}
-
-} // namespace Common

src/common/uint128.h

@@ -4,19 +4,118 @@
 
 #pragma once
 
+#include <cstring>
 #include <utility>
+
+#ifdef _MSC_VER
+#include <intrin.h>
+#pragma intrinsic(__umulh)
+#pragma intrinsic(_umul128)
+#pragma intrinsic(_udiv128)
+#else
+#include <x86intrin.h>
+#endif
+
 #include "common/common_types.h"
 
 namespace Common {
 
 // This function multiplies 2 u64 values and divides it by a u64 value.
-[[nodiscard]] u64 MultiplyAndDivide64(u64 a, u64 b, u64 d);
+[[nodiscard]] static inline u64 MultiplyAndDivide64(u64 a, u64 b, u64 d) {
+#ifdef _MSC_VER
+    u128 r{};
+    r[0] = _umul128(a, b, &r[1]);
+    u64 remainder;
+#if _MSC_VER < 1923
+    return udiv128(r[1], r[0], d, &remainder);
+#else
+    return _udiv128(r[1], r[0], d, &remainder);
+#endif
+#else
+    const u64 diva = a / d;
+    const u64 moda = a % d;
+    const u64 divb = b / d;
+    const u64 modb = b % d;
+    return diva * b + moda * divb + moda * modb / d;
+#endif
+}
 
 // This function multiplies 2 u64 values and produces a u128 value;
-[[nodiscard]] u128 Multiply64Into128(u64 a, u64 b);
+[[nodiscard]] static inline u128 Multiply64Into128(u64 a, u64 b) {
+    u128 result;
+#ifdef _MSC_VER
+    result[0] = _umul128(a, b, &result[1]);
+#else
+    unsigned __int128 tmp = a;
+    tmp *= b;
+    std::memcpy(&result, &tmp, sizeof(u128));
+#endif
+    return result;
+}
+
+[[nodiscard]] static inline u64 GetFixedPoint64Factor(u64 numerator, u64 divisor) {
+#ifdef __SIZEOF_INT128__
+    const auto base = static_cast<unsigned __int128>(numerator) << 64ULL;
+    return static_cast<u64>(base / divisor);
+#elif defined(_M_X64) || defined(_M_ARM64)
+    std::array<u64, 2> r = {0, numerator};
+    u64 remainder;
+#if _MSC_VER < 1923
+    return udiv128(r[1], r[0], divisor, &remainder);
+#else
+    return _udiv128(r[1], r[0], divisor, &remainder);
+#endif
+#else
+    // This one is bit more inaccurate.
+    return MultiplyAndDivide64(std::numeric_limits<u64>::max(), numerator, divisor);
+#endif
+}
+
+[[nodiscard]] static inline u64 MultiplyHigh(u64 a, u64 b) {
+#ifdef __SIZEOF_INT128__
+    return (static_cast<unsigned __int128>(a) * static_cast<unsigned __int128>(b)) >> 64;
+#elif defined(_M_X64) || defined(_M_ARM64)
+    return __umulh(a, b); // MSVC
+#else
+    // Generic fallback
+    const u64 a_lo = u32(a);
+    const u64 a_hi = a >> 32;
+    const u64 b_lo = u32(b);
+    const u64 b_hi = b >> 32;
+
+    const u64 a_x_b_hi = a_hi * b_hi;
+    const u64 a_x_b_mid = a_hi * b_lo;
+    const u64 b_x_a_mid = b_hi * a_lo;
+    const u64 a_x_b_lo = a_lo * b_lo;
+
+    const u64 carry_bit = (static_cast<u64>(static_cast<u32>(a_x_b_mid)) +
+                           static_cast<u64>(static_cast<u32>(b_x_a_mid)) + (a_x_b_lo >> 32)) >>
+                          32;
+
+    const u64 multhi = a_x_b_hi + (a_x_b_mid >> 32) + (b_x_a_mid >> 32) + carry_bit;
+
+    return multhi;
+#endif
+}
 
 // This function divides a u128 by a u32 value and produces two u64 values:
 // the result of division and the remainder
-[[nodiscard]] std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor);
+[[nodiscard]] static inline std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor) {
+    u64 remainder = dividend[0] % divisor;
+    u64 accum = dividend[0] / divisor;
+    if (dividend[1] == 0)
+        return {accum, remainder};
+    // We ignore dividend[1] / divisor as that overflows
+    const u64 first_segment = (dividend[1] % divisor) << 32;
+    accum += (first_segment / divisor) << 32;
+    const u64 second_segment = (first_segment % divisor) << 32;
+    accum += (second_segment / divisor);
+    remainder += second_segment % divisor;
+    if (remainder >= divisor) {
+        accum++;
+        remainder -= divisor;
+    }
+    return {accum, remainder};
+}
 
 } // namespace Common

src/common/wall_clock.cpp

@@ -2,6 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cstdint>
+
 #include "common/uint128.h"
 #include "common/wall_clock.h"
 

src/common/x64/native_clock.cpp

@@ -8,68 +8,10 @@
 #include <mutex>
 #include <thread>
 
-#ifdef _MSC_VER
-#include <intrin.h>
-
-#pragma intrinsic(__umulh)
-#pragma intrinsic(_udiv128)
-#else
-#include <x86intrin.h>
-#endif
-
 #include "common/atomic_ops.h"
 #include "common/uint128.h"
 #include "common/x64/native_clock.h"
 
-namespace {
-
-[[nodiscard]] u64 GetFixedPoint64Factor(u64 numerator, u64 divisor) {
-#ifdef __SIZEOF_INT128__
-    const auto base = static_cast<unsigned __int128>(numerator) << 64ULL;
-    return static_cast<u64>(base / divisor);
-#elif defined(_M_X64) || defined(_M_ARM64)
-    std::array<u64, 2> r = {0, numerator};
-    u64 remainder;
-#if _MSC_VER < 1923
-    return udiv128(r[1], r[0], divisor, &remainder);
-#else
-    return _udiv128(r[1], r[0], divisor, &remainder);
-#endif
-#else
-    // This one is bit more inaccurate.
-    return MultiplyAndDivide64(std::numeric_limits<u64>::max(), numerator, divisor);
-#endif
-}
-
-[[nodiscard]] u64 MultiplyHigh(u64 a, u64 b) {
-#ifdef __SIZEOF_INT128__
-    return (static_cast<unsigned __int128>(a) * static_cast<unsigned __int128>(b)) >> 64;
-#elif defined(_M_X64) || defined(_M_ARM64)
-    return __umulh(a, b); // MSVC
-#else
-    // Generic fallback
-    const u64 a_lo = u32(a);
-    const u64 a_hi = a >> 32;
-    const u64 b_lo = u32(b);
-    const u64 b_hi = b >> 32;
-
-    const u64 a_x_b_hi = a_hi * b_hi;
-    const u64 a_x_b_mid = a_hi * b_lo;
-    const u64 b_x_a_mid = b_hi * a_lo;
-    const u64 a_x_b_lo = a_lo * b_lo;
-
-    const u64 carry_bit = (static_cast<u64>(static_cast<u32>(a_x_b_mid)) +
-                           static_cast<u64>(static_cast<u32>(b_x_a_mid)) + (a_x_b_lo >> 32)) >>
-                          32;
-
-    const u64 multhi = a_x_b_hi + (a_x_b_mid >> 32) + (b_x_a_mid >> 32) + carry_bit;
-
-    return multhi;
-#endif
-}
-
-} // namespace
-
 namespace Common {
 
 u64 EstimateRDTSCFrequency() {

src/core/CMakeLists.txt

@@ -19,7 +19,6 @@ add_library(core STATIC
     core.h
     core_timing.cpp
     core_timing.h
-    core_timing_util.cpp
     core_timing_util.h
     cpu_manager.cpp
     cpu_manager.h
@@ -142,6 +141,9 @@ add_library(core STATIC
     hardware_interrupt_manager.h
     hle/ipc.h
     hle/ipc_helpers.h
+    hle/kernel/board/nintendo/nx/k_system_control.cpp
+    hle/kernel/board/nintendo/nx/k_system_control.h
+    hle/kernel/board/nintendo/nx/secure_monitor.h
     hle/kernel/client_port.cpp
     hle/kernel/client_port.h
     hle/kernel/client_session.cpp
@@ -157,6 +159,8 @@ add_library(core STATIC
     hle/kernel/hle_ipc.h
     hle/kernel/k_address_arbiter.cpp
     hle/kernel/k_address_arbiter.h
+    hle/kernel/k_address_space_info.cpp
+    hle/kernel/k_address_space_info.h
     hle/kernel/k_affinity_mask.h
     hle/kernel/k_condition_variable.cpp
     hle/kernel/k_condition_variable.h
@@ -165,6 +169,22 @@ add_library(core STATIC
     hle/kernel/k_light_condition_variable.h
     hle/kernel/k_light_lock.cpp
     hle/kernel/k_light_lock.h
+    hle/kernel/k_memory_block.h
+    hle/kernel/k_memory_block_manager.cpp
+    hle/kernel/k_memory_block_manager.h
+    hle/kernel/k_memory_layout.cpp
+    hle/kernel/k_memory_layout.board.nintendo_nx.cpp
+    hle/kernel/k_memory_layout.h
+    hle/kernel/k_memory_manager.cpp
+    hle/kernel/k_memory_manager.h
+    hle/kernel/k_memory_region.h
+    hle/kernel/k_memory_region_type.h
+    hle/kernel/k_page_bitmap.h
+    hle/kernel/k_page_heap.cpp
+    hle/kernel/k_page_heap.h
+    hle/kernel/k_page_linked_list.h
+    hle/kernel/k_page_table.cpp
+    hle/kernel/k_page_table.h
     hle/kernel/k_priority_queue.h
     hle/kernel/k_readable_event.cpp
     hle/kernel/k_readable_event.h
@@ -176,32 +196,23 @@ add_library(core STATIC
     hle/kernel/k_scoped_lock.h
     hle/kernel/k_scoped_resource_reservation.h
     hle/kernel/k_scoped_scheduler_lock_and_sleep.h
+    hle/kernel/k_shared_memory.cpp
+    hle/kernel/k_shared_memory.h
+    hle/kernel/k_slab_heap.h
+    hle/kernel/k_spin_lock.cpp
+    hle/kernel/k_spin_lock.h
     hle/kernel/k_synchronization_object.cpp
     hle/kernel/k_synchronization_object.h
+    hle/kernel/k_system_control.h
     hle/kernel/k_thread.cpp
     hle/kernel/k_thread.h
     hle/kernel/k_thread_queue.h
+    hle/kernel/k_trace.h
     hle/kernel/k_writable_event.cpp
     hle/kernel/k_writable_event.h
     hle/kernel/kernel.cpp
     hle/kernel/kernel.h
-    hle/kernel/memory/address_space_info.cpp
-    hle/kernel/memory/address_space_info.h
-    hle/kernel/memory/memory_block.h
-    hle/kernel/memory/memory_block_manager.cpp
-    hle/kernel/memory/memory_block_manager.h
-    hle/kernel/memory/memory_layout.h
-    hle/kernel/memory/memory_manager.cpp
-    hle/kernel/memory/memory_manager.h
-    hle/kernel/memory/memory_types.h
-    hle/kernel/memory/page_linked_list.h
-    hle/kernel/memory/page_heap.cpp
-    hle/kernel/memory/page_heap.h
-    hle/kernel/memory/page_table.cpp
-    hle/kernel/memory/page_table.h
-    hle/kernel/memory/slab_heap.h
-    hle/kernel/memory/system_control.cpp
-    hle/kernel/memory/system_control.h
+    hle/kernel/memory_types.h
     hle/kernel/object.cpp
     hle/kernel/object.h
     hle/kernel/physical_core.cpp
@@ -219,8 +230,6 @@ add_library(core STATIC
     hle/kernel/service_thread.h
     hle/kernel/session.cpp
     hle/kernel/session.h
-    hle/kernel/shared_memory.cpp
-    hle/kernel/shared_memory.h
     hle/kernel/svc.cpp
     hle/kernel/svc.h
     hle/kernel/svc_common.h
@@ -266,6 +275,7 @@ add_library(core STATIC
     hle/service/am/applets/software_keyboard.h
     hle/service/am/applets/web_browser.cpp
     hle/service/am/applets/web_browser.h
+    hle/service/am/applets/web_types.h
     hle/service/am/idle.cpp
     hle/service/am/idle.h
     hle/service/am/omm.cpp
@@ -400,6 +410,7 @@ add_library(core STATIC
     hle/service/hid/controllers/xpad.h
     hle/service/lbl/lbl.cpp
     hle/service/lbl/lbl.h
+    hle/service/ldn/errors.h
     hle/service/ldn/ldn.cpp
     hle/service/ldn/ldn.h
     hle/service/ldr/ldr.cpp
@@ -662,7 +673,7 @@ endif()
 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::fmt nlohmann_json::nlohmann_json mbedtls opus zip)
+target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt::fmt nlohmann_json::nlohmann_json mbedtls Opus::Opus zip)
 
 if (YUZU_ENABLE_BOXCAT)
     target_compile_definitions(core PRIVATE -DYUZU_ENABLE_BOXCAT)

src/core/arm/dynarmic/arm_dynarmic_32.cpp

@@ -114,18 +114,17 @@ public:
     static constexpr u64 minimum_run_cycles = 1000U;
 };
 
-std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable& page_table,
-                                                             std::size_t address_space_bits) const {
+std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable* page_table) const {
     Dynarmic::A32::UserConfig config;
     config.callbacks = cb.get();
-    // TODO(bunnei): Implement page table for 32-bit
-    // config.page_table = &page_table.pointers;
     config.coprocessors[15] = cp15;
     config.define_unpredictable_behaviour = true;
     static constexpr std::size_t PAGE_BITS = 12;
     static constexpr std::size_t NUM_PAGE_TABLE_ENTRIES = 1 << (32 - PAGE_BITS);
-    config.page_table = reinterpret_cast<std::array<std::uint8_t*, NUM_PAGE_TABLE_ENTRIES>*>(
-        page_table.pointers.data());
+    if (page_table) {
+        config.page_table = reinterpret_cast<std::array<std::uint8_t*, NUM_PAGE_TABLE_ENTRIES>*>(
+            page_table->pointers.data());
+    }
     config.absolute_offset_page_table = true;
     config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS;
     config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128;
@@ -138,6 +137,10 @@ std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable&
     // Timing
     config.wall_clock_cntpct = uses_wall_clock;
 
+    // Code cache size
+    config.code_cache_size = 512 * 1024 * 1024;
+    config.far_code_offset = 256 * 1024 * 1024;
+
     // Safe optimizations
     if (Settings::values.cpu_accuracy == Settings::CPUAccuracy::DebugMode) {
         if (!Settings::values.cpuopt_page_tables) {
@@ -201,7 +204,8 @@ ARM_Dynarmic_32::ARM_Dynarmic_32(System& system, CPUInterrupts& interrupt_handle
     : ARM_Interface{system, interrupt_handlers, uses_wall_clock},
       cb(std::make_unique<DynarmicCallbacks32>(*this)),
       cp15(std::make_shared<DynarmicCP15>(*this)), core_index{core_index},
-      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {}
+      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)},
+      jit(MakeJit(nullptr)) {}
 
 ARM_Dynarmic_32::~ARM_Dynarmic_32() = default;
 
@@ -256,9 +260,6 @@ void ARM_Dynarmic_32::ChangeProcessorID(std::size_t new_core_id) {
 }
 
 void ARM_Dynarmic_32::SaveContext(ThreadContext32& ctx) {
-    if (!jit) {
-        return;
-    }
     Dynarmic::A32::Context context;
     jit->SaveContext(context);
     ctx.cpu_registers = context.Regs();
@@ -268,9 +269,6 @@ void ARM_Dynarmic_32::SaveContext(ThreadContext32& ctx) {
 }
 
 void ARM_Dynarmic_32::LoadContext(const ThreadContext32& ctx) {
-    if (!jit) {
-        return;
-    }
     Dynarmic::A32::Context context;
     context.Regs() = ctx.cpu_registers;
     context.ExtRegs() = ctx.extension_registers;
@@ -284,35 +282,31 @@ void ARM_Dynarmic_32::PrepareReschedule() {
 }
 
 void ARM_Dynarmic_32::ClearInstructionCache() {
-    if (!jit) {
-        return;
-    }
     jit->ClearCache();
 }
 
 void ARM_Dynarmic_32::InvalidateCacheRange(VAddr addr, std::size_t size) {
-    if (!jit) {
-        return;
-    }
     jit->InvalidateCacheRange(static_cast<u32>(addr), size);
 }
 
 void ARM_Dynarmic_32::ClearExclusiveState() {
-    if (!jit) {
-        return;
-    }
     jit->ClearExclusiveState();
 }
 
 void ARM_Dynarmic_32::PageTableChanged(Common::PageTable& page_table,
                                        std::size_t new_address_space_size_in_bits) {
+    ThreadContext32 ctx{};
+    SaveContext(ctx);
+
     auto key = std::make_pair(&page_table, new_address_space_size_in_bits);
     auto iter = jit_cache.find(key);
     if (iter != jit_cache.end()) {
         jit = iter->second;
+        LoadContext(ctx);
         return;
     }
-    jit = MakeJit(page_table, new_address_space_size_in_bits);
+    jit = MakeJit(&page_table);
+    LoadContext(ctx);
     jit_cache.emplace(key, jit);
 }
 

src/core/arm/dynarmic/arm_dynarmic_32.h

@@ -68,8 +68,7 @@ public:
                           std::size_t new_address_space_size_in_bits) override;
 
 private:
-    std::shared_ptr<Dynarmic::A32::Jit> MakeJit(Common::PageTable& page_table,
-                                                std::size_t address_space_bits) const;
+    std::shared_ptr<Dynarmic::A32::Jit> MakeJit(Common::PageTable* page_table) const;
 
     using JitCacheKey = std::pair<Common::PageTable*, std::size_t>;
     using JitCacheType =
@@ -80,10 +79,10 @@ private:
 
     std::unique_ptr<DynarmicCallbacks32> cb;
     JitCacheType jit_cache;
-    std::shared_ptr<Dynarmic::A32::Jit> jit;
     std::shared_ptr<DynarmicCP15> cp15;
     std::size_t core_index;
     DynarmicExclusiveMonitor& exclusive_monitor;
+    std::shared_ptr<Dynarmic::A32::Jit> jit;
 };
 
 } // namespace Core

src/core/arm/dynarmic/arm_dynarmic_64.cpp

@@ -142,7 +142,7 @@ public:
     static constexpr u64 minimum_run_cycles = 1000U;
 };
 
-std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& page_table,
+std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable* page_table,
                                                              std::size_t address_space_bits) const {
     Dynarmic::A64::UserConfig config;
 
@@ -150,13 +150,15 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable&
     config.callbacks = cb.get();
 
     // Memory
-    config.page_table = reinterpret_cast<void**>(page_table.pointers.data());
-    config.page_table_address_space_bits = address_space_bits;
-    config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS;
-    config.silently_mirror_page_table = false;
-    config.absolute_offset_page_table = true;
-    config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128;
-    config.only_detect_misalignment_via_page_table_on_page_boundary = true;
+    if (page_table) {
+        config.page_table = reinterpret_cast<void**>(page_table->pointers.data());
+        config.page_table_address_space_bits = address_space_bits;
+        config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS;
+        config.silently_mirror_page_table = false;
+        config.absolute_offset_page_table = true;
+        config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128;
+        config.only_detect_misalignment_via_page_table_on_page_boundary = true;
+    }
 
     // Multi-process state
     config.processor_id = core_index;
@@ -175,6 +177,10 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable&
     // Timing
     config.wall_clock_cntpct = uses_wall_clock;
 
+    // Code cache size
+    config.code_cache_size = 512 * 1024 * 1024;
+    config.far_code_offset = 256 * 1024 * 1024;
+
     // Safe optimizations
     if (Settings::values.cpu_accuracy == Settings::CPUAccuracy::DebugMode) {
         if (!Settings::values.cpuopt_page_tables) {
@@ -237,7 +243,8 @@ ARM_Dynarmic_64::ARM_Dynarmic_64(System& system, CPUInterrupts& interrupt_handle
                                  std::size_t core_index)
     : ARM_Interface{system, interrupt_handlers, uses_wall_clock},
       cb(std::make_unique<DynarmicCallbacks64>(*this)), core_index{core_index},
-      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {}
+      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)},
+      jit(MakeJit(nullptr, 48)) {}
 
 ARM_Dynarmic_64::~ARM_Dynarmic_64() = default;
 
@@ -294,9 +301,6 @@ void ARM_Dynarmic_64::ChangeProcessorID(std::size_t new_core_id) {
 }
 
 void ARM_Dynarmic_64::SaveContext(ThreadContext64& ctx) {
-    if (!jit) {
-        return;
-    }
     ctx.cpu_registers = jit->GetRegisters();
     ctx.sp = jit->GetSP();
     ctx.pc = jit->GetPC();
@@ -308,9 +312,6 @@ void ARM_Dynarmic_64::SaveContext(ThreadContext64& ctx) {
 }
 
 void ARM_Dynarmic_64::LoadContext(const ThreadContext64& ctx) {
-    if (!jit) {
-        return;
-    }
     jit->SetRegisters(ctx.cpu_registers);
     jit->SetSP(ctx.sp);
     jit->SetPC(ctx.pc);
@@ -326,35 +327,31 @@ void ARM_Dynarmic_64::PrepareReschedule() {
 }
 
 void ARM_Dynarmic_64::ClearInstructionCache() {
-    if (!jit) {
-        return;
-    }
     jit->ClearCache();
 }
 
 void ARM_Dynarmic_64::InvalidateCacheRange(VAddr addr, std::size_t size) {
-    if (!jit) {
-        return;
-    }
     jit->InvalidateCacheRange(addr, size);
 }
 
 void ARM_Dynarmic_64::ClearExclusiveState() {
-    if (!jit) {
-        return;
-    }
     jit->ClearExclusiveState();
 }
 
 void ARM_Dynarmic_64::PageTableChanged(Common::PageTable& page_table,
                                        std::size_t new_address_space_size_in_bits) {
+    ThreadContext64 ctx{};
+    SaveContext(ctx);
+
     auto key = std::make_pair(&page_table, new_address_space_size_in_bits);
     auto iter = jit_cache.find(key);
     if (iter != jit_cache.end()) {
         jit = iter->second;
+        LoadContext(ctx);
         return;
     }
-    jit = MakeJit(page_table, new_address_space_size_in_bits);
+    jit = MakeJit(&page_table, new_address_space_size_in_bits);
+    LoadContext(ctx);
     jit_cache.emplace(key, jit);
 }
 

src/core/arm/dynarmic/arm_dynarmic_64.h

@@ -61,7 +61,7 @@ public:
                           std::size_t new_address_space_size_in_bits) override;
 
 private:
-    std::shared_ptr<Dynarmic::A64::Jit> MakeJit(Common::PageTable& page_table,
+    std::shared_ptr<Dynarmic::A64::Jit> MakeJit(Common::PageTable* page_table,
                                                 std::size_t address_space_bits) const;
 
     using JitCacheKey = std::pair<Common::PageTable*, std::size_t>;
@@ -71,10 +71,11 @@ private:
     friend class DynarmicCallbacks64;
     std::unique_ptr<DynarmicCallbacks64> cb;
     JitCacheType jit_cache;
-    std::shared_ptr<Dynarmic::A64::Jit> jit;
 
     std::size_t core_index;
     DynarmicExclusiveMonitor& exclusive_monitor;
+
+    std::shared_ptr<Dynarmic::A64::Jit> jit;
 };
 
 } // namespace Core

src/core/core.cpp

@@ -296,28 +296,20 @@ struct System::Impl {
         exit_lock = false;
 
         if (gpu_core) {
-            gpu_core->WaitIdle();
+            gpu_core->ShutDown();
         }
 
-        // Shutdown emulation session
         services.reset();
         service_manager.reset();
         cheat_engine.reset();
         telemetry_session.reset();
-
-        // Close all CPU/threading state
         cpu_manager.Shutdown();
-
-        // Shutdown kernel and core timing
+        time_manager.Shutdown();
         core_timing.Shutdown();
-        kernel.Shutdown();
-
-        // Close app loader
         app_loader.reset();
         gpu_core.reset();
         perf_stats.reset();
-
-        // Clear all applets
+        kernel.Shutdown();
         applet_manager.ClearAll();
 
         LOG_DEBUG(Core, "Shutdown OK");

src/core/core_timing_util.cpp

@@ -1,84 +0,0 @@
-// Copyright 2008 Dolphin Emulator Project / 2017 Citra Emulator Project
-// Licensed under GPLv2+
-// Refer to the license.txt file included.
-
-#include "core/core_timing_util.h"
-
-#include <cinttypes>
-#include <limits>
-#include "common/logging/log.h"
-#include "common/uint128.h"
-#include "core/hardware_properties.h"
-
-namespace Core::Timing {
-
-constexpr u64 MAX_VALUE_TO_MULTIPLY = std::numeric_limits<s64>::max() / Hardware::BASE_CLOCK_RATE;
-
-s64 msToCycles(std::chrono::milliseconds ms) {
-    if (static_cast<u64>(ms.count() / 1000) > MAX_VALUE_TO_MULTIPLY) {
-        LOG_ERROR(Core_Timing, "Integer overflow, use max value");
-        return std::numeric_limits<s64>::max();
-    }
-    if (static_cast<u64>(ms.count()) > MAX_VALUE_TO_MULTIPLY) {
-        LOG_DEBUG(Core_Timing, "Time very big, do rounding");
-        return Hardware::BASE_CLOCK_RATE * (ms.count() / 1000);
-    }
-    return (Hardware::BASE_CLOCK_RATE * ms.count()) / 1000;
-}
-
-s64 usToCycles(std::chrono::microseconds us) {
-    if (static_cast<u64>(us.count() / 1000000) > MAX_VALUE_TO_MULTIPLY) {
-        LOG_ERROR(Core_Timing, "Integer overflow, use max value");
-        return std::numeric_limits<s64>::max();
-    }
-    if (static_cast<u64>(us.count()) > MAX_VALUE_TO_MULTIPLY) {
-        LOG_DEBUG(Core_Timing, "Time very big, do rounding");
-        return Hardware::BASE_CLOCK_RATE * (us.count() / 1000000);
-    }
-    return (Hardware::BASE_CLOCK_RATE * us.count()) / 1000000;
-}
-
-s64 nsToCycles(std::chrono::nanoseconds ns) {
-    const u128 temporal = Common::Multiply64Into128(ns.count(), Hardware::BASE_CLOCK_RATE);
-    return Common::Divide128On32(temporal, static_cast<u32>(1000000000)).first;
-}
-
-u64 msToClockCycles(std::chrono::milliseconds ns) {
-    const u128 temp = Common::Multiply64Into128(ns.count(), Hardware::CNTFREQ);
-    return Common::Divide128On32(temp, 1000).first;
-}
-
-u64 usToClockCycles(std::chrono::microseconds ns) {
-    const u128 temp = Common::Multiply64Into128(ns.count(), Hardware::CNTFREQ);
-    return Common::Divide128On32(temp, 1000000).first;
-}
-
-u64 nsToClockCycles(std::chrono::nanoseconds ns) {
-    const u128 temp = Common::Multiply64Into128(ns.count(), Hardware::CNTFREQ);
-    return Common::Divide128On32(temp, 1000000000).first;
-}
-
-u64 CpuCyclesToClockCycles(u64 ticks) {
-    const u128 temporal = Common::Multiply64Into128(ticks, Hardware::CNTFREQ);
-    return Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first;
-}
-
-std::chrono::milliseconds CyclesToMs(s64 cycles) {
-    const u128 temporal = Common::Multiply64Into128(cycles, 1000);
-    u64 ms = Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first;
-    return std::chrono::milliseconds(ms);
-}
-
-std::chrono::nanoseconds CyclesToNs(s64 cycles) {
-    const u128 temporal = Common::Multiply64Into128(cycles, 1000000000);
-    u64 ns = Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first;
-    return std::chrono::nanoseconds(ns);
-}
-
-std::chrono::microseconds CyclesToUs(s64 cycles) {
-    const u128 temporal = Common::Multiply64Into128(cycles, 1000000);
-    u64 us = Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first;
-    return std::chrono::microseconds(us);
-}
-
-} // namespace Core::Timing

src/core/core_timing_util.h

@@ -1,24 +1,59 @@
-// Copyright 2008 Dolphin Emulator Project / 2017 Citra Emulator Project
-// Licensed under GPLv2+
+// Copyright 2020 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
 #pragma once
 
 #include <chrono>
+
 #include "common/common_types.h"
+#include "core/hardware_properties.h"
 
 namespace Core::Timing {
 
-s64 msToCycles(std::chrono::milliseconds ms);
-s64 usToCycles(std::chrono::microseconds us);
-s64 nsToCycles(std::chrono::nanoseconds ns);
-u64 msToClockCycles(std::chrono::milliseconds ns);
-u64 usToClockCycles(std::chrono::microseconds ns);
-u64 nsToClockCycles(std::chrono::nanoseconds ns);
-std::chrono::milliseconds CyclesToMs(s64 cycles);
-std::chrono::nanoseconds CyclesToNs(s64 cycles);
-std::chrono::microseconds CyclesToUs(s64 cycles);
+namespace detail {
+constexpr u64 CNTFREQ_ADJUSTED = Hardware::CNTFREQ / 1000;
+constexpr u64 BASE_CLOCK_RATE_ADJUSTED = Hardware::BASE_CLOCK_RATE / 1000;
+} // namespace detail
 
-u64 CpuCyclesToClockCycles(u64 ticks);
+[[nodiscard]] constexpr s64 msToCycles(std::chrono::milliseconds ms) {
+    return ms.count() * detail::BASE_CLOCK_RATE_ADJUSTED;
+}
+
+[[nodiscard]] constexpr s64 usToCycles(std::chrono::microseconds us) {
+    return us.count() * detail::BASE_CLOCK_RATE_ADJUSTED / 1000;
+}
+
+[[nodiscard]] constexpr s64 nsToCycles(std::chrono::nanoseconds ns) {
+    return ns.count() * detail::BASE_CLOCK_RATE_ADJUSTED / 1000000;
+}
+
+[[nodiscard]] constexpr u64 msToClockCycles(std::chrono::milliseconds ms) {
+    return static_cast<u64>(ms.count()) * detail::CNTFREQ_ADJUSTED;
+}
+
+[[nodiscard]] constexpr u64 usToClockCycles(std::chrono::microseconds us) {
+    return us.count() * detail::CNTFREQ_ADJUSTED / 1000;
+}
+
+[[nodiscard]] constexpr u64 nsToClockCycles(std::chrono::nanoseconds ns) {
+    return ns.count() * detail::CNTFREQ_ADJUSTED / 1000000;
+}
+
+[[nodiscard]] constexpr u64 CpuCyclesToClockCycles(u64 ticks) {
+    return ticks * detail::CNTFREQ_ADJUSTED / detail::BASE_CLOCK_RATE_ADJUSTED;
+}
+
+[[nodiscard]] constexpr std::chrono::milliseconds CyclesToMs(s64 cycles) {
+    return std::chrono::milliseconds(cycles / detail::BASE_CLOCK_RATE_ADJUSTED);
+}
+
+[[nodiscard]] constexpr std::chrono::nanoseconds CyclesToNs(s64 cycles) {
+    return std::chrono::nanoseconds(cycles * 1000000 / detail::BASE_CLOCK_RATE_ADJUSTED);
+}
+
+[[nodiscard]] constexpr std::chrono::microseconds CyclesToUs(s64 cycles) {
+    return std::chrono::microseconds(cycles * 1000 / detail::BASE_CLOCK_RATE_ADJUSTED);
+}
 
 } // namespace Core::Timing

src/core/cpu_manager.cpp

@@ -148,7 +148,7 @@ void CpuManager::MultiCoreRunSuspendThread() {
         auto core = kernel.GetCurrentHostThreadID();
         auto& scheduler = *kernel.CurrentScheduler();
         Kernel::KThread* current_thread = scheduler.GetCurrentThread();
-        Common::Fiber::YieldTo(current_thread->GetHostContext(), core_data[core].host_context);
+        Common::Fiber::YieldTo(current_thread->GetHostContext(), *core_data[core].host_context);
         ASSERT(scheduler.ContextSwitchPending());
         ASSERT(core == kernel.GetCurrentHostThreadID());
         scheduler.RescheduleCurrentCore();
@@ -245,7 +245,7 @@ void CpuManager::SingleCoreRunSuspendThread() {
         auto core = kernel.GetCurrentHostThreadID();
         auto& scheduler = *kernel.CurrentScheduler();
         Kernel::KThread* current_thread = scheduler.GetCurrentThread();
-        Common::Fiber::YieldTo(current_thread->GetHostContext(), core_data[0].host_context);
+        Common::Fiber::YieldTo(current_thread->GetHostContext(), *core_data[0].host_context);
         ASSERT(scheduler.ContextSwitchPending());
         ASSERT(core == kernel.GetCurrentHostThreadID());
         scheduler.RescheduleCurrentCore();
@@ -271,7 +271,7 @@ void CpuManager::PreemptSingleCore(bool from_running_enviroment) {
         scheduler.Unload(scheduler.GetCurrentThread());
 
         auto& next_scheduler = kernel.Scheduler(current_core);
-        Common::Fiber::YieldTo(current_thread->GetHostContext(), next_scheduler.ControlContext());
+        Common::Fiber::YieldTo(current_thread->GetHostContext(), *next_scheduler.ControlContext());
     }
 
     // May have changed scheduler
@@ -363,7 +363,7 @@ void CpuManager::RunThread(std::size_t core) {
 
         auto current_thread = system.Kernel().CurrentScheduler()->GetCurrentThread();
         data.is_running = true;
-        Common::Fiber::YieldTo(data.host_context, current_thread->GetHostContext());
+        Common::Fiber::YieldTo(data.host_context, *current_thread->GetHostContext());
         data.is_running = false;
         data.is_paused = true;
         data.exit_barrier->Wait();

src/core/crypto/aes_util.cpp

@@ -105,8 +105,6 @@ void AESCipher<Key, KeySize>::Transcode(const u8* src, std::size_t size, u8* des
             }
         }
     }
-
-    mbedtls_cipher_finish(context, nullptr, nullptr);
 }
 
 template <typename Key, std::size_t KeySize>

src/core/file_sys/control_metadata.cpp

@@ -100,6 +100,14 @@ u64 NACP::GetDeviceSaveDataSize() const {
     return raw.device_save_data_size;
 }
 
+u32 NACP::GetParentalControlFlag() const {
+    return raw.parental_control;
+}
+
+const std::array<u8, 0x20>& NACP::GetRatingAge() const {
+    return raw.rating_age;
+}
+
 std::vector<u8> NACP::GetRawBytes() const {
     std::vector<u8> out(sizeof(RawNACP));
     std::memcpy(out.data(), &raw, sizeof(RawNACP));

src/core/file_sys/control_metadata.h

@@ -114,6 +114,8 @@ public:
     std::vector<u8> GetRawBytes() const;
     bool GetUserAccountSwitchLock() const;
     u64 GetDeviceSaveDataSize() const;
+    u32 GetParentalControlFlag() const;
+    const std::array<u8, 0x20>& GetRatingAge() const;
 
 private:
     RawNACP raw{};

src/core/file_sys/system_archive/ng_word.cpp

@@ -14,7 +14,7 @@ namespace NgWord1Data {
 constexpr std::size_t NUMBER_WORD_TXT_FILES = 0x10;
 
 // Should this archive replacement mysteriously not work on a future game, consider updating.
-constexpr std::array<u8, 4> VERSION_DAT{0x0, 0x0, 0x0, 0x19}; // 5.1.0 System Version
+constexpr std::array<u8, 4> VERSION_DAT{0x0, 0x0, 0x0, 0x20}; // 11.0.1 System Version
 
 constexpr std::array<u8, 30> WORD_TXT{
     0xFE, 0xFF, 0x00, 0x5E, 0x00, 0x76, 0x00, 0x65, 0x00, 0x72, 0x00, 0x79, 0x00, 0x62, 0x00,
@@ -43,7 +43,7 @@ namespace NgWord2Data {
 constexpr std::size_t NUMBER_AC_NX_FILES = 0x10;
 
 // Should this archive replacement mysteriously not work on a future game, consider updating.
-constexpr std::array<u8, 4> VERSION_DAT{0x0, 0x0, 0x0, 0x15}; // 5.1.0 System Version
+constexpr std::array<u8, 4> VERSION_DAT{0x0, 0x0, 0x0, 0x1A}; // 11.0.1 System Version
 
 constexpr std::array<u8, 0x2C> AC_NX_DATA{
     0x1F, 0x8B, 0x08, 0x08, 0xD5, 0x2C, 0x09, 0x5C, 0x04, 0x00, 0x61, 0x63, 0x72, 0x61, 0x77,

src/core/file_sys/system_archive/system_version.cpp

@@ -14,15 +14,15 @@ namespace SystemVersionData {
 
 constexpr u8 VERSION_MAJOR = 11;
 constexpr u8 VERSION_MINOR = 0;
-constexpr u8 VERSION_MICRO = 0;
+constexpr u8 VERSION_MICRO = 1;
 
-constexpr u8 REVISION_MAJOR = 5;
+constexpr u8 REVISION_MAJOR = 1;
 constexpr u8 REVISION_MINOR = 0;
 
 constexpr char PLATFORM_STRING[] = "NX";
-constexpr char VERSION_HASH[] = "34197eba8810e2edd5e9dfcfbde7b340882e856d";
-constexpr char DISPLAY_VERSION[] = "11.0.0";
-constexpr char DISPLAY_TITLE[] = "NintendoSDK Firmware for NX 11.0.0-5.0";
+constexpr char VERSION_HASH[] = "69103fcb2004dace877094c2f8c29e6113be5dbf";
+constexpr char DISPLAY_VERSION[] = "11.0.1";
+constexpr char DISPLAY_TITLE[] = "NintendoSDK Firmware for NX 11.0.1-1.0";
 
 } // namespace SystemVersionData
 

src/core/frontend/applets/controller.h

@@ -31,6 +31,7 @@ struct ControllerParameters {
     bool allow_dual_joycons{};
     bool allow_left_joycon{};
     bool allow_right_joycon{};
+    bool allow_gamecube_controller{};
 };
 
 class ControllerApplet {

src/core/hle/kernel/arch/arm64/k_memory_region_device_types.inc

@@ -0,0 +1,20 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+// All architectures must define NumArchitectureDeviceRegions.
+constexpr inline const auto NumArchitectureDeviceRegions = 3;
+
+constexpr inline const auto KMemoryRegionType_Uart =
+    KMemoryRegionType_ArchDeviceBase.DeriveSparse(0, NumArchitectureDeviceRegions, 0);
+constexpr inline const auto KMemoryRegionType_InterruptCpuInterface =
+    KMemoryRegionType_ArchDeviceBase.DeriveSparse(0, NumArchitectureDeviceRegions, 1)
+        .SetAttribute(KMemoryRegionAttr_NoUserMap);
+constexpr inline const auto KMemoryRegionType_InterruptDistributor =
+    KMemoryRegionType_ArchDeviceBase.DeriveSparse(0, NumArchitectureDeviceRegions, 2)
+        .SetAttribute(KMemoryRegionAttr_NoUserMap);
+static_assert(KMemoryRegionType_Uart.GetValue() == (0x1D));
+static_assert(KMemoryRegionType_InterruptCpuInterface.GetValue() ==
+              (0x2D | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_InterruptDistributor.GetValue() ==
+              (0x4D | KMemoryRegionAttr_NoUserMap));

src/core/hle/kernel/board/nintendo/nx/k_memory_region_device_types.inc

@@ -0,0 +1,52 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+// All architectures must define NumBoardDeviceRegions.
+constexpr inline const auto NumBoardDeviceRegions = 6;
+// UNUSED: .Derive(NumBoardDeviceRegions, 0);
+constexpr inline const auto KMemoryRegionType_MemoryController =
+    KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 1)
+        .SetAttribute(KMemoryRegionAttr_NoUserMap);
+constexpr inline const auto KMemoryRegionType_MemoryController1 =
+    KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 2)
+        .SetAttribute(KMemoryRegionAttr_NoUserMap);
+constexpr inline const auto KMemoryRegionType_MemoryController0 =
+    KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 3)
+        .SetAttribute(KMemoryRegionAttr_NoUserMap);
+constexpr inline const auto KMemoryRegionType_PowerManagementController =
+    KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 4).DeriveTransition();
+constexpr inline const auto KMemoryRegionType_LegacyLpsDevices =
+    KMemoryRegionType_BoardDeviceBase.Derive(NumBoardDeviceRegions, 5);
+static_assert(KMemoryRegionType_MemoryController.GetValue() ==
+              (0x55 | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_MemoryController1.GetValue() ==
+              (0x65 | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_MemoryController0.GetValue() ==
+              (0x95 | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_PowerManagementController.GetValue() == (0x1A5));
+
+static_assert(KMemoryRegionType_LegacyLpsDevices.GetValue() == 0xC5);
+
+constexpr inline const auto NumLegacyLpsDevices = 7;
+constexpr inline const auto KMemoryRegionType_LegacyLpsExceptionVectors =
+    KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 0);
+constexpr inline const auto KMemoryRegionType_LegacyLpsIram =
+    KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 1);
+constexpr inline const auto KMemoryRegionType_LegacyLpsFlowController =
+    KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 2);
+constexpr inline const auto KMemoryRegionType_LegacyLpsPrimaryICtlr =
+    KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 3);
+constexpr inline const auto KMemoryRegionType_LegacyLpsSemaphore =
+    KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 4);
+constexpr inline const auto KMemoryRegionType_LegacyLpsAtomics =
+    KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 5);
+constexpr inline const auto KMemoryRegionType_LegacyLpsClkRst =
+    KMemoryRegionType_LegacyLpsDevices.Derive(NumLegacyLpsDevices, 6);
+static_assert(KMemoryRegionType_LegacyLpsExceptionVectors.GetValue() == 0x3C5);
+static_assert(KMemoryRegionType_LegacyLpsIram.GetValue() == 0x5C5);
+static_assert(KMemoryRegionType_LegacyLpsFlowController.GetValue() == 0x6C5);
+static_assert(KMemoryRegionType_LegacyLpsPrimaryICtlr.GetValue() == 0x9C5);
+static_assert(KMemoryRegionType_LegacyLpsSemaphore.GetValue() == 0xAC5);
+static_assert(KMemoryRegionType_LegacyLpsAtomics.GetValue() == 0xCC5);
+static_assert(KMemoryRegionType_LegacyLpsClkRst.GetValue() == 0x11C5);

src/core/hle/kernel/board/nintendo/nx/k_system_control.cpp

@@ -0,0 +1,164 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <random>
+
+#include "common/common_sizes.h"
+#include "core/hle/kernel/board/nintendo/nx/k_system_control.h"
+#include "core/hle/kernel/board/nintendo/nx/secure_monitor.h"
+#include "core/hle/kernel/k_trace.h"
+
+namespace Kernel::Board::Nintendo::Nx {
+
+namespace impl {
+
+constexpr const std::size_t RequiredNonSecureSystemMemorySizeVi = 0x2238 * 4 * 1024;
+constexpr const std::size_t RequiredNonSecureSystemMemorySizeNvservices = 0x710 * 4 * 1024;
+constexpr const std::size_t RequiredNonSecureSystemMemorySizeMisc = 0x80 * 4 * 1024;
+
+} // namespace impl
+
+constexpr const std::size_t RequiredNonSecureSystemMemorySize =
+    impl::RequiredNonSecureSystemMemorySizeVi + impl::RequiredNonSecureSystemMemorySizeNvservices +
+    impl::RequiredNonSecureSystemMemorySizeMisc;
+
+namespace {
+
+u32 GetMemoryModeForInit() {
+    return 0x01;
+}
+
+u32 GetMemorySizeForInit() {
+    return 0;
+}
+
+Smc::MemoryArrangement GetMemoryArrangeForInit() {
+    switch (GetMemoryModeForInit() & 0x3F) {
+    case 0x01:
+    default:
+        return Smc::MemoryArrangement_4GB;
+    case 0x02:
+        return Smc::MemoryArrangement_4GBForAppletDev;
+    case 0x03:
+        return Smc::MemoryArrangement_4GBForSystemDev;
+    case 0x11:
+        return Smc::MemoryArrangement_6GB;
+    case 0x12:
+        return Smc::MemoryArrangement_6GBForAppletDev;
+    case 0x21:
+        return Smc::MemoryArrangement_8GB;
+    }
+}
+} // namespace
+
+// Initialization.
+size_t KSystemControl::Init::GetIntendedMemorySize() {
+    switch (GetMemorySizeForInit()) {
+    case Smc::MemorySize_4GB:
+    default: // All invalid modes should go to 4GB.
+        return Common::Size_4_GB;
+    case Smc::MemorySize_6GB:
+        return Common::Size_6_GB;
+    case Smc::MemorySize_8GB:
+        return Common::Size_8_GB;
+    }
+}
+
+PAddr KSystemControl::Init::GetKernelPhysicalBaseAddress(u64 base_address) {
+    return base_address;
+}
+
+bool KSystemControl::Init::ShouldIncreaseThreadResourceLimit() {
+    return true;
+}
+
+std::size_t KSystemControl::Init::GetApplicationPoolSize() {
+    // Get the base pool size.
+    const size_t base_pool_size = []() -> size_t {
+        switch (GetMemoryArrangeForInit()) {
+        case Smc::MemoryArrangement_4GB:
+        default:
+            return Common::Size_3285_MB;
+        case Smc::MemoryArrangement_4GBForAppletDev:
+            return Common::Size_2048_MB;
+        case Smc::MemoryArrangement_4GBForSystemDev:
+            return Common::Size_3285_MB;
+        case Smc::MemoryArrangement_6GB:
+            return Common::Size_4916_MB;
+        case Smc::MemoryArrangement_6GBForAppletDev:
+            return Common::Size_3285_MB;
+        case Smc::MemoryArrangement_8GB:
+            return Common::Size_4916_MB;
+        }
+    }();
+
+    // Return (possibly) adjusted size.
+    return base_pool_size;
+}
+
+size_t KSystemControl::Init::GetAppletPoolSize() {
+    // Get the base pool size.
+    const size_t base_pool_size = []() -> size_t {
+        switch (GetMemoryArrangeForInit()) {
+        case Smc::MemoryArrangement_4GB:
+        default:
+            return Common::Size_507_MB;
+        case Smc::MemoryArrangement_4GBForAppletDev:
+            return Common::Size_1554_MB;
+        case Smc::MemoryArrangement_4GBForSystemDev:
+            return Common::Size_448_MB;
+        case Smc::MemoryArrangement_6GB:
+            return Common::Size_562_MB;
+        case Smc::MemoryArrangement_6GBForAppletDev:
+            return Common::Size_2193_MB;
+        case Smc::MemoryArrangement_8GB:
+            return Common::Size_2193_MB;
+        }
+    }();
+
+    // Return (possibly) adjusted size.
+    constexpr size_t ExtraSystemMemoryForAtmosphere = Common::Size_33_MB;
+    return base_pool_size - ExtraSystemMemoryForAtmosphere - KTraceBufferSize;
+}
+
+size_t KSystemControl::Init::GetMinimumNonSecureSystemPoolSize() {
+    // Verify that our minimum is at least as large as Nintendo's.
+    constexpr size_t MinimumSize = RequiredNonSecureSystemMemorySize;
+    static_assert(MinimumSize >= 0x29C8000);
+
+    return MinimumSize;
+}
+
+namespace {
+template <typename F>
+u64 GenerateUniformRange(u64 min, u64 max, F f) {
+    // Handle the case where the difference is too large to represent.
+    if (max == std::numeric_limits<u64>::max() && min == std::numeric_limits<u64>::min()) {
+        return f();
+    }
+
+    // Iterate until we get a value in range.
+    const u64 range_size = ((max + 1) - min);
+    const u64 effective_max = (std::numeric_limits<u64>::max() / range_size) * range_size;
+    while (true) {
+        if (const u64 rnd = f(); rnd < effective_max) {
+            return min + (rnd % range_size);
+        }
+    }
+}
+
+} // Anonymous namespace
+
+u64 KSystemControl::GenerateRandomU64() {
+    static std::random_device device;
+    static std::mt19937 gen(device());
+    static std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max());
+    return distribution(gen);
+}
+
+u64 KSystemControl::GenerateRandomRange(u64 min, u64 max) {
+    return GenerateUniformRange(min, max, GenerateRandomU64);
+}
+
+} // namespace Kernel::Board::Nintendo::Nx

src/core/hle/kernel/board/nintendo/nx/k_system_control.h

@@ -0,0 +1,28 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+
+namespace Kernel::Board::Nintendo::Nx {
+
+class KSystemControl {
+public:
+    class Init {
+    public:
+        // Initialization.
+        static std::size_t GetIntendedMemorySize();
+        static PAddr GetKernelPhysicalBaseAddress(u64 base_address);
+        static bool ShouldIncreaseThreadResourceLimit();
+        static std::size_t GetApplicationPoolSize();
+        static std::size_t GetAppletPoolSize();
+        static std::size_t GetMinimumNonSecureSystemPoolSize();
+    };
+
+    static u64 GenerateRandomRange(u64 min, u64 max);
+    static u64 GenerateRandomU64();
+};
+
+} // namespace Kernel::Board::Nintendo::Nx

src/core/hle/kernel/board/nintendo/nx/secure_monitor.h

@@ -0,0 +1,26 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+
+namespace Kernel::Board::Nintendo::Nx::Smc {
+
+enum MemorySize {
+    MemorySize_4GB = 0,
+    MemorySize_6GB = 1,
+    MemorySize_8GB = 2,
+};
+
+enum MemoryArrangement {
+    MemoryArrangement_4GB = 0,
+    MemoryArrangement_4GBForAppletDev = 1,
+    MemoryArrangement_4GBForSystemDev = 2,
+    MemoryArrangement_6GB = 3,
+    MemoryArrangement_6GBForAppletDev = 4,
+    MemoryArrangement_8GB = 5,
+};
+
+} // namespace Kernel::Board::Nintendo::Nx::Smc

src/core/hle/kernel/k_address_space_info.cpp

@@ -0,0 +1,106 @@
+// Copyright 2020 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <array>
+
+#include "common/assert.h"
+#include "common/common_sizes.h"
+#include "core/hle/kernel/k_address_space_info.h"
+
+namespace Kernel {
+
+namespace {
+
+// clang-format off
+constexpr std::array<KAddressSpaceInfo, 13> AddressSpaceInfos{{
+   { .bit_width = 32, .address = Common::Size_2_MB   , .size = Common::Size_1_GB   - Common::Size_2_MB  , .type = KAddressSpaceInfo::Type::MapSmall, },
+   { .bit_width = 32, .address = Common::Size_1_GB   , .size = Common::Size_4_GB   - Common::Size_1_GB  , .type = KAddressSpaceInfo::Type::MapLarge, },
+   { .bit_width = 32, .address = Common::Size_Invalid, .size = Common::Size_1_GB                        , .type = KAddressSpaceInfo::Type::Alias,    },
+   { .bit_width = 32, .address = Common::Size_Invalid, .size = Common::Size_1_GB                        , .type = KAddressSpaceInfo::Type::Heap,     },
+   { .bit_width = 36, .address = Common::Size_128_MB , .size = Common::Size_2_GB   - Common::Size_128_MB, .type = KAddressSpaceInfo::Type::MapSmall, },
+   { .bit_width = 36, .address = Common::Size_2_GB   , .size = Common::Size_64_GB  - Common::Size_2_GB  , .type = KAddressSpaceInfo::Type::MapLarge, },
+   { .bit_width = 36, .address = Common::Size_Invalid, .size = Common::Size_6_GB                        , .type = KAddressSpaceInfo::Type::Heap,     },
+   { .bit_width = 36, .address = Common::Size_Invalid, .size = Common::Size_6_GB                        , .type = KAddressSpaceInfo::Type::Alias,    },
+   { .bit_width = 39, .address = Common::Size_128_MB , .size = Common::Size_512_GB - Common::Size_128_MB, .type = KAddressSpaceInfo::Type::Map39Bit, },
+   { .bit_width = 39, .address = Common::Size_Invalid, .size = Common::Size_64_GB                       , .type = KAddressSpaceInfo::Type::MapSmall  },
+   { .bit_width = 39, .address = Common::Size_Invalid, .size = Common::Size_6_GB                        , .type = KAddressSpaceInfo::Type::Heap,     },
+   { .bit_width = 39, .address = Common::Size_Invalid, .size = Common::Size_64_GB                       , .type = KAddressSpaceInfo::Type::Alias,    },
+   { .bit_width = 39, .address = Common::Size_Invalid, .size = Common::Size_2_GB                        , .type = KAddressSpaceInfo::Type::Stack,    },
+}};
+// clang-format on
+
+constexpr bool IsAllowedIndexForAddress(std::size_t index) {
+    return index < AddressSpaceInfos.size() &&
+           AddressSpaceInfos[index].address != Common::Size_Invalid;
+}
+
+using IndexArray =
+    std::array<std::size_t, static_cast<std::size_t>(KAddressSpaceInfo::Type::Count)>;
+
+constexpr IndexArray AddressSpaceIndices32Bit{
+    0, 1, 0, 2, 0, 3,
+};
+
+constexpr IndexArray AddressSpaceIndices36Bit{
+    4, 5, 4, 6, 4, 7,
+};
+
+constexpr IndexArray AddressSpaceIndices39Bit{
+    9, 8, 8, 10, 12, 11,
+};
+
+constexpr bool IsAllowed32BitType(KAddressSpaceInfo::Type type) {
+    return type < KAddressSpaceInfo::Type::Count && type != KAddressSpaceInfo::Type::Map39Bit &&
+           type != KAddressSpaceInfo::Type::Stack;
+}
+
+constexpr bool IsAllowed36BitType(KAddressSpaceInfo::Type type) {
+    return type < KAddressSpaceInfo::Type::Count && type != KAddressSpaceInfo::Type::Map39Bit &&
+           type != KAddressSpaceInfo::Type::Stack;
+}
+
+constexpr bool IsAllowed39BitType(KAddressSpaceInfo::Type type) {
+    return type < KAddressSpaceInfo::Type::Count && type != KAddressSpaceInfo::Type::MapLarge;
+}
+
+} // namespace
+
+u64 KAddressSpaceInfo::GetAddressSpaceStart(std::size_t width, Type type) {
+    const std::size_t index{static_cast<std::size_t>(type)};
+    switch (width) {
+    case 32:
+        ASSERT(IsAllowed32BitType(type));
+        ASSERT(IsAllowedIndexForAddress(AddressSpaceIndices32Bit[index]));
+        return AddressSpaceInfos[AddressSpaceIndices32Bit[index]].address;
+    case 36:
+        ASSERT(IsAllowed36BitType(type));
+        ASSERT(IsAllowedIndexForAddress(AddressSpaceIndices36Bit[index]));
+        return AddressSpaceInfos[AddressSpaceIndices36Bit[index]].address;
+    case 39:
+        ASSERT(IsAllowed39BitType(type));
+        ASSERT(IsAllowedIndexForAddress(AddressSpaceIndices39Bit[index]));
+        return AddressSpaceInfos[AddressSpaceIndices39Bit[index]].address;
+    }
+    UNREACHABLE();
+    return 0;
+}
+
+std::size_t KAddressSpaceInfo::GetAddressSpaceSize(std::size_t width, Type type) {
+    const std::size_t index{static_cast<std::size_t>(type)};
+    switch (width) {
+    case 32:
+        ASSERT(IsAllowed32BitType(type));
+        return AddressSpaceInfos[AddressSpaceIndices32Bit[index]].size;
+    case 36:
+        ASSERT(IsAllowed36BitType(type));
+        return AddressSpaceInfos[AddressSpaceIndices36Bit[index]].size;
+    case 39:
+        ASSERT(IsAllowed39BitType(type));
+        return AddressSpaceInfos[AddressSpaceIndices39Bit[index]].size;
+    }
+    UNREACHABLE();
+    return 0;
+}
+
+} // namespace Kernel

src/core/hle/kernel/k_address_space_info.h

@@ -2,20 +2,17 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-// This file references various implementation details from Atmosphere, an open-source firmware for
-// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX.
-
 #pragma once
 
 #include "common/common_types.h"
 
-namespace Kernel::Memory {
+namespace Kernel {
 
-struct AddressSpaceInfo final {
+struct KAddressSpaceInfo final {
     enum class Type : u32 {
-        Is32Bit = 0,
-        Small64Bit = 1,
-        Large64Bit = 2,
+        MapSmall = 0,
+        MapLarge = 1,
+        Map39Bit = 2,
         Heap = 3,
         Stack = 4,
         Alias = 5,
@@ -31,4 +28,4 @@ struct AddressSpaceInfo final {
     const Type type{};
 };
 
-} // namespace Kernel::Memory
+} // namespace Kernel

src/core/hle/kernel/k_memory_block.h

@@ -2,20 +2,17 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-// This file references various implementation details from Atmosphere, an open-source firmware for
-// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX.
-
 #pragma once
 
 #include "common/alignment.h"
 #include "common/assert.h"
 #include "common/common_types.h"
-#include "core/hle/kernel/memory/memory_types.h"
+#include "core/hle/kernel/memory_types.h"
 #include "core/hle/kernel/svc_types.h"
 
-namespace Kernel::Memory {
+namespace Kernel {
 
-enum class MemoryState : u32 {
+enum class KMemoryState : u32 {
     None = 0,
     Mask = 0xFF,
     All = ~None,
@@ -97,31 +94,31 @@ enum class MemoryState : u32 {
                     FlagReferenceCounted | FlagCanDebug,
     CodeOut = static_cast<u32>(Svc::MemoryState::CodeOut) | FlagMapped | FlagReferenceCounted,
 };
-DECLARE_ENUM_FLAG_OPERATORS(MemoryState);
+DECLARE_ENUM_FLAG_OPERATORS(KMemoryState);
 
-static_assert(static_cast<u32>(MemoryState::Free) == 0x00000000);
-static_assert(static_cast<u32>(MemoryState::Io) == 0x00002001);
-static_assert(static_cast<u32>(MemoryState::Static) == 0x00042002);
-static_assert(static_cast<u32>(MemoryState::Code) == 0x00DC7E03);
-static_assert(static_cast<u32>(MemoryState::CodeData) == 0x03FEBD04);
-static_assert(static_cast<u32>(MemoryState::Normal) == 0x037EBD05);
-static_assert(static_cast<u32>(MemoryState::Shared) == 0x00402006);
-static_assert(static_cast<u32>(MemoryState::AliasCode) == 0x00DD7E08);
-static_assert(static_cast<u32>(MemoryState::AliasCodeData) == 0x03FFBD09);
-static_assert(static_cast<u32>(MemoryState::Ipc) == 0x005C3C0A);
-static_assert(static_cast<u32>(MemoryState::Stack) == 0x005C3C0B);
-static_assert(static_cast<u32>(MemoryState::ThreadLocal) == 0x0040200C);
-static_assert(static_cast<u32>(MemoryState::Transferred) == 0x015C3C0D);
-static_assert(static_cast<u32>(MemoryState::SharedTransferred) == 0x005C380E);
-static_assert(static_cast<u32>(MemoryState::SharedCode) == 0x0040380F);
-static_assert(static_cast<u32>(MemoryState::Inaccessible) == 0x00000010);
-static_assert(static_cast<u32>(MemoryState::NonSecureIpc) == 0x005C3811);
-static_assert(static_cast<u32>(MemoryState::NonDeviceIpc) == 0x004C2812);
-static_assert(static_cast<u32>(MemoryState::Kernel) == 0x00002013);
-static_assert(static_cast<u32>(MemoryState::GeneratedCode) == 0x00402214);
-static_assert(static_cast<u32>(MemoryState::CodeOut) == 0x00402015);
+static_assert(static_cast<u32>(KMemoryState::Free) == 0x00000000);
+static_assert(static_cast<u32>(KMemoryState::Io) == 0x00002001);
+static_assert(static_cast<u32>(KMemoryState::Static) == 0x00042002);
+static_assert(static_cast<u32>(KMemoryState::Code) == 0x00DC7E03);
+static_assert(static_cast<u32>(KMemoryState::CodeData) == 0x03FEBD04);
+static_assert(static_cast<u32>(KMemoryState::Normal) == 0x037EBD05);
+static_assert(static_cast<u32>(KMemoryState::Shared) == 0x00402006);
+static_assert(static_cast<u32>(KMemoryState::AliasCode) == 0x00DD7E08);
+static_assert(static_cast<u32>(KMemoryState::AliasCodeData) == 0x03FFBD09);
+static_assert(static_cast<u32>(KMemoryState::Ipc) == 0x005C3C0A);
+static_assert(static_cast<u32>(KMemoryState::Stack) == 0x005C3C0B);
+static_assert(static_cast<u32>(KMemoryState::ThreadLocal) == 0x0040200C);
+static_assert(static_cast<u32>(KMemoryState::Transferred) == 0x015C3C0D);
+static_assert(static_cast<u32>(KMemoryState::SharedTransferred) == 0x005C380E);
+static_assert(static_cast<u32>(KMemoryState::SharedCode) == 0x0040380F);
+static_assert(static_cast<u32>(KMemoryState::Inaccessible) == 0x00000010);
+static_assert(static_cast<u32>(KMemoryState::NonSecureIpc) == 0x005C3811);
+static_assert(static_cast<u32>(KMemoryState::NonDeviceIpc) == 0x004C2812);
+static_assert(static_cast<u32>(KMemoryState::Kernel) == 0x00002013);
+static_assert(static_cast<u32>(KMemoryState::GeneratedCode) == 0x00402214);
+static_assert(static_cast<u32>(KMemoryState::CodeOut) == 0x00402015);
 
-enum class MemoryPermission : u8 {
+enum class KMemoryPermission : u8 {
     None = 0,
     Mask = static_cast<u8>(~None),
 
@@ -135,9 +132,9 @@ enum class MemoryPermission : u8 {
     UserMask = static_cast<u8>(Svc::MemoryPermission::Read | Svc::MemoryPermission::Write |
                                Svc::MemoryPermission::Execute),
 };
-DECLARE_ENUM_FLAG_OPERATORS(MemoryPermission);
+DECLARE_ENUM_FLAG_OPERATORS(KMemoryPermission);
 
-enum class MemoryAttribute : u8 {
+enum class KMemoryAttribute : u8 {
     None = 0x00,
     Mask = 0x7F,
     All = Mask,
@@ -152,18 +149,18 @@ enum class MemoryAttribute : u8 {
     LockedAndIpcLocked = Locked | IpcLocked,
     DeviceSharedAndUncached = DeviceShared | Uncached
 };
-DECLARE_ENUM_FLAG_OPERATORS(MemoryAttribute);
+DECLARE_ENUM_FLAG_OPERATORS(KMemoryAttribute);
 
-static_assert((static_cast<u8>(MemoryAttribute::Mask) &
-               static_cast<u8>(MemoryAttribute::DontCareMask)) == 0);
+static_assert((static_cast<u8>(KMemoryAttribute::Mask) &
+               static_cast<u8>(KMemoryAttribute::DontCareMask)) == 0);
 
-struct MemoryInfo {
+struct KMemoryInfo {
     VAddr addr{};
     std::size_t size{};
-    MemoryState state{};
-    MemoryPermission perm{};
-    MemoryAttribute attribute{};
-    MemoryPermission original_perm{};
+    KMemoryState state{};
+    KMemoryPermission perm{};
+    KMemoryAttribute attribute{};
+    KMemoryPermission original_perm{};
     u16 ipc_lock_count{};
     u16 device_use_count{};
 
@@ -171,9 +168,9 @@ struct MemoryInfo {
         return {
             addr,
             size,
-            static_cast<Svc::MemoryState>(state & MemoryState::Mask),
-            static_cast<Svc::MemoryAttribute>(attribute & MemoryAttribute::Mask),
-            static_cast<Svc::MemoryPermission>(perm & MemoryPermission::UserMask),
+            static_cast<Svc::MemoryState>(state & KMemoryState::Mask),
+            static_cast<Svc::MemoryAttribute>(attribute & KMemoryAttribute::Mask),
+            static_cast<Svc::MemoryPermission>(perm & KMemoryPermission::UserMask),
             ipc_lock_count,
             device_use_count,
         };
@@ -196,21 +193,21 @@ struct MemoryInfo {
     }
 };
 
-class MemoryBlock final {
-    friend class MemoryBlockManager;
+class KMemoryBlock final {
+    friend class KMemoryBlockManager;
 
 private:
     VAddr addr{};
     std::size_t num_pages{};
-    MemoryState state{MemoryState::None};
+    KMemoryState state{KMemoryState::None};
     u16 ipc_lock_count{};
     u16 device_use_count{};
-    MemoryPermission perm{MemoryPermission::None};
-    MemoryPermission original_perm{MemoryPermission::None};
-    MemoryAttribute attribute{MemoryAttribute::None};
+    KMemoryPermission perm{KMemoryPermission::None};
+    KMemoryPermission original_perm{KMemoryPermission::None};
+    KMemoryAttribute attribute{KMemoryAttribute::None};
 
 public:
-    static constexpr int Compare(const MemoryBlock& lhs, const MemoryBlock& rhs) {
+    static constexpr int Compare(const KMemoryBlock& lhs, const KMemoryBlock& rhs) {
         if (lhs.GetAddress() < rhs.GetAddress()) {
             return -1;
         } else if (lhs.GetAddress() <= rhs.GetLastAddress()) {
@@ -221,9 +218,9 @@ public:
     }
 
 public:
-    constexpr MemoryBlock() = default;
-    constexpr MemoryBlock(VAddr addr_, std::size_t num_pages_, MemoryState state_,
-                          MemoryPermission perm_, MemoryAttribute attribute_)
+    constexpr KMemoryBlock() = default;
+    constexpr KMemoryBlock(VAddr addr_, std::size_t num_pages_, KMemoryState state_,
+                           KMemoryPermission perm_, KMemoryAttribute attribute_)
         : addr{addr_}, num_pages(num_pages_), state{state_}, perm{perm_}, attribute{attribute_} {}
 
     constexpr VAddr GetAddress() const {
@@ -246,40 +243,40 @@ public:
         return GetEndAddress() - 1;
     }
 
-    constexpr MemoryInfo GetMemoryInfo() const {
+    constexpr KMemoryInfo GetMemoryInfo() const {
         return {
             GetAddress(), GetSize(),     state,          perm,
             attribute,    original_perm, ipc_lock_count, device_use_count,
         };
     }
 
-    void ShareToDevice(MemoryPermission /*new_perm*/) {
-        ASSERT((attribute & MemoryAttribute::DeviceShared) == MemoryAttribute::DeviceShared ||
+    void ShareToDevice(KMemoryPermission /*new_perm*/) {
+        ASSERT((attribute & KMemoryAttribute::DeviceShared) == KMemoryAttribute::DeviceShared ||
                device_use_count == 0);
-        attribute |= MemoryAttribute::DeviceShared;
+        attribute |= KMemoryAttribute::DeviceShared;
         const u16 new_use_count{++device_use_count};
         ASSERT(new_use_count > 0);
     }
 
-    void UnshareToDevice(MemoryPermission /*new_perm*/) {
-        ASSERT((attribute & MemoryAttribute::DeviceShared) == MemoryAttribute::DeviceShared);
+    void UnshareToDevice(KMemoryPermission /*new_perm*/) {
+        ASSERT((attribute & KMemoryAttribute::DeviceShared) == KMemoryAttribute::DeviceShared);
         const u16 prev_use_count{device_use_count--};
         ASSERT(prev_use_count > 0);
         if (prev_use_count == 1) {
-            attribute &= ~MemoryAttribute::DeviceShared;
+            attribute &= ~KMemoryAttribute::DeviceShared;
         }
     }
 
 private:
-    constexpr bool HasProperties(MemoryState s, MemoryPermission p, MemoryAttribute a) const {
-        constexpr MemoryAttribute AttributeIgnoreMask{MemoryAttribute::DontCareMask |
-                                                      MemoryAttribute::IpcLocked |
-                                                      MemoryAttribute::DeviceShared};
+    constexpr bool HasProperties(KMemoryState s, KMemoryPermission p, KMemoryAttribute a) const {
+        constexpr KMemoryAttribute AttributeIgnoreMask{KMemoryAttribute::DontCareMask |
+                                                       KMemoryAttribute::IpcLocked |
+                                                       KMemoryAttribute::DeviceShared};
         return state == s && perm == p &&
                (attribute | AttributeIgnoreMask) == (a | AttributeIgnoreMask);
     }
 
-    constexpr bool HasSameProperties(const MemoryBlock& rhs) const {
+    constexpr bool HasSameProperties(const KMemoryBlock& rhs) const {
         return state == rhs.state && perm == rhs.perm && original_perm == rhs.original_perm &&
                attribute == rhs.attribute && ipc_lock_count == rhs.ipc_lock_count &&
                device_use_count == rhs.device_use_count;
@@ -296,25 +293,25 @@ private:
         num_pages += count;
     }
 
-    constexpr void Update(MemoryState new_state, MemoryPermission new_perm,
-                          MemoryAttribute new_attribute) {
-        ASSERT(original_perm == MemoryPermission::None);
-        ASSERT((attribute & MemoryAttribute::IpcLocked) == MemoryAttribute::None);
+    constexpr void Update(KMemoryState new_state, KMemoryPermission new_perm,
+                          KMemoryAttribute new_attribute) {
+        ASSERT(original_perm == KMemoryPermission::None);
+        ASSERT((attribute & KMemoryAttribute::IpcLocked) == KMemoryAttribute::None);
 
         state = new_state;
         perm = new_perm;
 
-        attribute = static_cast<MemoryAttribute>(
+        attribute = static_cast<KMemoryAttribute>(
             new_attribute |
-            (attribute & (MemoryAttribute::IpcLocked | MemoryAttribute::DeviceShared)));
+            (attribute & (KMemoryAttribute::IpcLocked | KMemoryAttribute::DeviceShared)));
     }
 
-    constexpr MemoryBlock Split(VAddr split_addr) {
+    constexpr KMemoryBlock Split(VAddr split_addr) {
         ASSERT(GetAddress() < split_addr);
         ASSERT(Contains(split_addr));
         ASSERT(Common::IsAligned(split_addr, PageSize));
 
-        MemoryBlock block;
+        KMemoryBlock block;
         block.addr = addr;
         block.num_pages = (split_addr - GetAddress()) / PageSize;
         block.state = state;
@@ -330,6 +327,6 @@ private:
         return block;
     }
 };
-static_assert(std::is_trivially_destructible<MemoryBlock>::value);
+static_assert(std::is_trivially_destructible<KMemoryBlock>::value);
 
-} // namespace Kernel::Memory
+} // namespace Kernel

src/core/hle/kernel/k_memory_block_manager.cpp

@@ -2,19 +2,19 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include "core/hle/kernel/memory/memory_block_manager.h"
-#include "core/hle/kernel/memory/memory_types.h"
+#include "core/hle/kernel/k_memory_block_manager.h"
+#include "core/hle/kernel/memory_types.h"
 
-namespace Kernel::Memory {
+namespace Kernel {
 
-MemoryBlockManager::MemoryBlockManager(VAddr start_addr, VAddr end_addr)
+KMemoryBlockManager::KMemoryBlockManager(VAddr start_addr, VAddr end_addr)
     : start_addr{start_addr}, end_addr{end_addr} {
     const u64 num_pages{(end_addr - start_addr) / PageSize};
-    memory_block_tree.emplace_back(start_addr, num_pages, MemoryState::Free, MemoryPermission::None,
-                                   MemoryAttribute::None);
+    memory_block_tree.emplace_back(start_addr, num_pages, KMemoryState::Free,
+                                   KMemoryPermission::None, KMemoryAttribute::None);
 }
 
-MemoryBlockManager::iterator MemoryBlockManager::FindIterator(VAddr addr) {
+KMemoryBlockManager::iterator KMemoryBlockManager::FindIterator(VAddr addr) {
     auto node{memory_block_tree.begin()};
     while (node != end()) {
         const VAddr end_addr{node->GetNumPages() * PageSize + node->GetAddress()};
@@ -26,9 +26,9 @@ MemoryBlockManager::iterator MemoryBlockManager::FindIterator(VAddr addr) {
     return end();
 }
 
-VAddr MemoryBlockManager::FindFreeArea(VAddr region_start, std::size_t region_num_pages,
-                                       std::size_t num_pages, std::size_t align, std::size_t offset,
-                                       std::size_t guard_pages) {
+VAddr KMemoryBlockManager::FindFreeArea(VAddr region_start, std::size_t region_num_pages,
+                                        std::size_t num_pages, std::size_t align,
+                                        std::size_t offset, std::size_t guard_pages) {
     if (num_pages == 0) {
         return {};
     }
@@ -41,7 +41,7 @@ VAddr MemoryBlockManager::FindFreeArea(VAddr region_start, std::size_t region_nu
             break;
         }
 
-        if (info.state != MemoryState::Free) {
+        if (info.state != KMemoryState::Free) {
             continue;
         }
 
@@ -63,17 +63,17 @@ VAddr MemoryBlockManager::FindFreeArea(VAddr region_start, std::size_t region_nu
     return {};
 }
 
-void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState prev_state,
-                                MemoryPermission prev_perm, MemoryAttribute prev_attribute,
-                                MemoryState state, MemoryPermission perm,
-                                MemoryAttribute attribute) {
+void KMemoryBlockManager::Update(VAddr addr, std::size_t num_pages, KMemoryState prev_state,
+                                 KMemoryPermission prev_perm, KMemoryAttribute prev_attribute,
+                                 KMemoryState state, KMemoryPermission perm,
+                                 KMemoryAttribute attribute) {
     const VAddr end_addr{addr + num_pages * PageSize};
     iterator node{memory_block_tree.begin()};
 
-    prev_attribute |= MemoryAttribute::IpcAndDeviceMapped;
+    prev_attribute |= KMemoryAttribute::IpcAndDeviceMapped;
 
     while (node != memory_block_tree.end()) {
-        MemoryBlock* block{&(*node)};
+        KMemoryBlock* block{&(*node)};
         iterator next_node{std::next(node)};
         const VAddr cur_addr{block->GetAddress()};
         const VAddr cur_end_addr{block->GetNumPages() * PageSize + cur_addr};
@@ -106,13 +106,13 @@ void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState p
     }
 }
 
-void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState state,
-                                MemoryPermission perm, MemoryAttribute attribute) {
+void KMemoryBlockManager::Update(VAddr addr, std::size_t num_pages, KMemoryState state,
+                                 KMemoryPermission perm, KMemoryAttribute attribute) {
     const VAddr end_addr{addr + num_pages * PageSize};
     iterator node{memory_block_tree.begin()};
 
     while (node != memory_block_tree.end()) {
-        MemoryBlock* block{&(*node)};
+        KMemoryBlock* block{&(*node)};
         iterator next_node{std::next(node)};
         const VAddr cur_addr{block->GetAddress()};
         const VAddr cur_end_addr{block->GetNumPages() * PageSize + cur_addr};
@@ -141,13 +141,13 @@ void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState s
     }
 }
 
-void MemoryBlockManager::UpdateLock(VAddr addr, std::size_t num_pages, LockFunc&& lock_func,
-                                    MemoryPermission perm) {
+void KMemoryBlockManager::UpdateLock(VAddr addr, std::size_t num_pages, LockFunc&& lock_func,
+                                     KMemoryPermission perm) {
     const VAddr end_addr{addr + num_pages * PageSize};
     iterator node{memory_block_tree.begin()};
 
     while (node != memory_block_tree.end()) {
-        MemoryBlock* block{&(*node)};
+        KMemoryBlock* block{&(*node)};
         iterator next_node{std::next(node)};
         const VAddr cur_addr{block->GetAddress()};
         const VAddr cur_end_addr{block->GetNumPages() * PageSize + cur_addr};
@@ -176,9 +176,9 @@ void MemoryBlockManager::UpdateLock(VAddr addr, std::size_t num_pages, LockFunc&
     }
 }
 
-void MemoryBlockManager::IterateForRange(VAddr start, VAddr end, IterateFunc&& func) {
+void KMemoryBlockManager::IterateForRange(VAddr start, VAddr end, IterateFunc&& func) {
     const_iterator it{FindIterator(start)};
-    MemoryInfo info{};
+    KMemoryInfo info{};
     do {
         info = it->GetMemoryInfo();
         func(info);
@@ -186,8 +186,8 @@ void MemoryBlockManager::IterateForRange(VAddr start, VAddr end, IterateFunc&& f
     } while (info.addr + info.size - 1 < end - 1 && it != cend());
 }
 
-void MemoryBlockManager::MergeAdjacent(iterator it, iterator& next_it) {
-    MemoryBlock* block{&(*it)};
+void KMemoryBlockManager::MergeAdjacent(iterator it, iterator& next_it) {
+    KMemoryBlock* block{&(*it)};
 
     auto EraseIt = [&](const iterator it_to_erase) {
         if (next_it == it_to_erase) {
@@ -197,7 +197,7 @@ void MemoryBlockManager::MergeAdjacent(iterator it, iterator& next_it) {
     };
 
     if (it != memory_block_tree.begin()) {
-        MemoryBlock* prev{&(*std::prev(it))};
+        KMemoryBlock* prev{&(*std::prev(it))};
 
         if (block->HasSameProperties(*prev)) {
             const iterator prev_it{std::prev(it)};
@@ -211,7 +211,7 @@ void MemoryBlockManager::MergeAdjacent(iterator it, iterator& next_it) {
     }
 
     if (it != cend()) {
-        const MemoryBlock* const next{&(*std::next(it))};
+        const KMemoryBlock* const next{&(*std::next(it))};
 
         if (block->HasSameProperties(*next)) {
             block->Add(next->GetNumPages());
@@ -220,4 +220,4 @@ void MemoryBlockManager::MergeAdjacent(iterator it, iterator& next_it) {
     }
 }
 
-} // namespace Kernel::Memory
+} // namespace Kernel

src/core/hle/kernel/k_memory_block_manager.h

@@ -8,18 +8,18 @@
 #include <list>
 
 #include "common/common_types.h"
-#include "core/hle/kernel/memory/memory_block.h"
+#include "core/hle/kernel/k_memory_block.h"
 
-namespace Kernel::Memory {
+namespace Kernel {
 
-class MemoryBlockManager final {
+class KMemoryBlockManager final {
 public:
-    using MemoryBlockTree = std::list<MemoryBlock>;
+    using MemoryBlockTree = std::list<KMemoryBlock>;
     using iterator = MemoryBlockTree::iterator;
     using const_iterator = MemoryBlockTree::const_iterator;
 
 public:
-    MemoryBlockManager(VAddr start_addr, VAddr end_addr);
+    KMemoryBlockManager(VAddr start_addr, VAddr end_addr);
 
     iterator end() {
         return memory_block_tree.end();
@@ -36,21 +36,22 @@ public:
     VAddr FindFreeArea(VAddr region_start, std::size_t region_num_pages, std::size_t num_pages,
                        std::size_t align, std::size_t offset, std::size_t guard_pages);
 
-    void Update(VAddr addr, std::size_t num_pages, MemoryState prev_state,
-                MemoryPermission prev_perm, MemoryAttribute prev_attribute, MemoryState state,
-                MemoryPermission perm, MemoryAttribute attribute);
+    void Update(VAddr addr, std::size_t num_pages, KMemoryState prev_state,
+                KMemoryPermission prev_perm, KMemoryAttribute prev_attribute, KMemoryState state,
+                KMemoryPermission perm, KMemoryAttribute attribute);
 
-    void Update(VAddr addr, std::size_t num_pages, MemoryState state,
-                MemoryPermission perm = MemoryPermission::None,
-                MemoryAttribute attribute = MemoryAttribute::None);
+    void Update(VAddr addr, std::size_t num_pages, KMemoryState state,
+                KMemoryPermission perm = KMemoryPermission::None,
+                KMemoryAttribute attribute = KMemoryAttribute::None);
 
-    using LockFunc = std::function<void(iterator, MemoryPermission)>;
-    void UpdateLock(VAddr addr, std::size_t num_pages, LockFunc&& lock_func, MemoryPermission perm);
+    using LockFunc = std::function<void(iterator, KMemoryPermission)>;
+    void UpdateLock(VAddr addr, std::size_t num_pages, LockFunc&& lock_func,
+                    KMemoryPermission perm);
 
-    using IterateFunc = std::function<void(const MemoryInfo&)>;
+    using IterateFunc = std::function<void(const KMemoryInfo&)>;
     void IterateForRange(VAddr start, VAddr end, IterateFunc&& func);
 
-    MemoryBlock& FindBlock(VAddr addr) {
+    KMemoryBlock& FindBlock(VAddr addr) {
         return *FindIterator(addr);
     }
 
@@ -63,4 +64,4 @@ private:
     MemoryBlockTree memory_block_tree;
 };
 
-} // namespace Kernel::Memory
+} // namespace Kernel

src/core/hle/kernel/k_memory_layout.board.nintendo_nx.cpp

@@ -0,0 +1,199 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/alignment.h"
+#include "core/hle/kernel/k_memory_layout.h"
+#include "core/hle/kernel/k_memory_manager.h"
+#include "core/hle/kernel/k_system_control.h"
+#include "core/hle/kernel/k_trace.h"
+
+namespace Kernel {
+
+namespace {
+
+constexpr size_t CarveoutAlignment = 0x20000;
+constexpr size_t CarveoutSizeMax = (512ULL * 1024 * 1024) - CarveoutAlignment;
+
+bool SetupPowerManagementControllerMemoryRegion(KMemoryLayout& memory_layout) {
+    // Above firmware 2.0.0, the PMC is not mappable.
+    return memory_layout.GetPhysicalMemoryRegionTree().Insert(
+               0x7000E000, 0x400, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap) &&
+           memory_layout.GetPhysicalMemoryRegionTree().Insert(
+               0x7000E400, 0xC00,
+               KMemoryRegionType_PowerManagementController | KMemoryRegionAttr_NoUserMap);
+}
+
+void InsertPoolPartitionRegionIntoBothTrees(KMemoryLayout& memory_layout, size_t start, size_t size,
+                                            KMemoryRegionType phys_type,
+                                            KMemoryRegionType virt_type, u32& cur_attr) {
+    const u32 attr = cur_attr++;
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(start, size,
+                                                              static_cast<u32>(phys_type), attr));
+    const KMemoryRegion* phys = memory_layout.GetPhysicalMemoryRegionTree().FindByTypeAndAttribute(
+        static_cast<u32>(phys_type), attr);
+    ASSERT(phys != nullptr);
+    ASSERT(phys->GetEndAddress() != 0);
+    ASSERT(memory_layout.GetVirtualMemoryRegionTree().Insert(phys->GetPairAddress(), size,
+                                                             static_cast<u32>(virt_type), attr));
+}
+
+} // namespace
+
+namespace Init {
+
+void SetupDevicePhysicalMemoryRegions(KMemoryLayout& memory_layout) {
+    ASSERT(SetupPowerManagementControllerMemoryRegion(memory_layout));
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x70019000, 0x1000, KMemoryRegionType_MemoryController | KMemoryRegionAttr_NoUserMap));
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x7001C000, 0x1000, KMemoryRegionType_MemoryController0 | KMemoryRegionAttr_NoUserMap));
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x7001D000, 0x1000, KMemoryRegionType_MemoryController1 | KMemoryRegionAttr_NoUserMap));
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x50040000, 0x1000, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap));
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x50041000, 0x1000,
+        KMemoryRegionType_InterruptDistributor | KMemoryRegionAttr_ShouldKernelMap));
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x50042000, 0x1000,
+        KMemoryRegionType_InterruptCpuInterface | KMemoryRegionAttr_ShouldKernelMap));
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x50043000, 0x1D000, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap));
+
+    // Map IRAM unconditionally, to support debug-logging-to-iram build config.
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x40000000, 0x40000, KMemoryRegionType_LegacyLpsIram | KMemoryRegionAttr_ShouldKernelMap));
+
+    // Above firmware 2.0.0, prevent mapping the bpmp exception vectors or the ipatch region.
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x6000F000, 0x1000, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap));
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        0x6001DC00, 0x400, KMemoryRegionType_None | KMemoryRegionAttr_NoUserMap));
+}
+
+void SetupDramPhysicalMemoryRegions(KMemoryLayout& memory_layout) {
+    const size_t intended_memory_size = KSystemControl::Init::GetIntendedMemorySize();
+    const PAddr physical_memory_base_address =
+        KSystemControl::Init::GetKernelPhysicalBaseAddress(DramPhysicalAddress);
+
+    // Insert blocks into the tree.
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        physical_memory_base_address, intended_memory_size, KMemoryRegionType_Dram));
+    ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+        physical_memory_base_address, ReservedEarlyDramSize, KMemoryRegionType_DramReservedEarly));
+
+    // Insert the KTrace block at the end of Dram, if KTrace is enabled.
+    static_assert(!IsKTraceEnabled || KTraceBufferSize > 0);
+    if constexpr (IsKTraceEnabled) {
+        const PAddr ktrace_buffer_phys_addr =
+            physical_memory_base_address + intended_memory_size - KTraceBufferSize;
+        ASSERT(memory_layout.GetPhysicalMemoryRegionTree().Insert(
+            ktrace_buffer_phys_addr, KTraceBufferSize, KMemoryRegionType_KernelTraceBuffer));
+    }
+}
+
+void SetupPoolPartitionMemoryRegions(KMemoryLayout& memory_layout) {
+    // Start by identifying the extents of the DRAM memory region.
+    const auto dram_extents = memory_layout.GetMainMemoryPhysicalExtents();
+    ASSERT(dram_extents.GetEndAddress() != 0);
+
+    // Determine the end of the pool region.
+    const u64 pool_end = dram_extents.GetEndAddress() - KTraceBufferSize;
+
+    // Find the start of the kernel DRAM region.
+    const KMemoryRegion* kernel_dram_region =
+        memory_layout.GetPhysicalMemoryRegionTree().FindFirstDerived(
+            KMemoryRegionType_DramKernelBase);
+    ASSERT(kernel_dram_region != nullptr);
+
+    const u64 kernel_dram_start = kernel_dram_region->GetAddress();
+    ASSERT(Common::IsAligned(kernel_dram_start, CarveoutAlignment));
+
+    // Find the start of the pool partitions region.
+    const KMemoryRegion* pool_partitions_region =
+        memory_layout.GetPhysicalMemoryRegionTree().FindByTypeAndAttribute(
+            KMemoryRegionType_DramPoolPartition, 0);
+    ASSERT(pool_partitions_region != nullptr);
+    const u64 pool_partitions_start = pool_partitions_region->GetAddress();
+
+    // Setup the pool partition layouts.
+    // On 5.0.0+, setup modern 4-pool-partition layout.
+
+    // Get Application and Applet pool sizes.
+    const size_t application_pool_size = KSystemControl::Init::GetApplicationPoolSize();
+    const size_t applet_pool_size = KSystemControl::Init::GetAppletPoolSize();
+    const size_t unsafe_system_pool_min_size =
+        KSystemControl::Init::GetMinimumNonSecureSystemPoolSize();
+
+    // Decide on starting addresses for our pools.
+    const u64 application_pool_start = pool_end - application_pool_size;
+    const u64 applet_pool_start = application_pool_start - applet_pool_size;
+    const u64 unsafe_system_pool_start = std::min(
+        kernel_dram_start + CarveoutSizeMax,
+        Common::AlignDown(applet_pool_start - unsafe_system_pool_min_size, CarveoutAlignment));
+    const size_t unsafe_system_pool_size = applet_pool_start - unsafe_system_pool_start;
+
+    // We want to arrange application pool depending on where the middle of dram is.
+    const u64 dram_midpoint = (dram_extents.GetAddress() + dram_extents.GetEndAddress()) / 2;
+    u32 cur_pool_attr = 0;
+    size_t total_overhead_size = 0;
+    if (dram_extents.GetEndAddress() <= dram_midpoint || dram_midpoint <= application_pool_start) {
+        InsertPoolPartitionRegionIntoBothTrees(
+            memory_layout, application_pool_start, application_pool_size,
+            KMemoryRegionType_DramApplicationPool, KMemoryRegionType_VirtualDramApplicationPool,
+            cur_pool_attr);
+        total_overhead_size +=
+            KMemoryManager::CalculateManagementOverheadSize(application_pool_size);
+    } else {
+        const size_t first_application_pool_size = dram_midpoint - application_pool_start;
+        const size_t second_application_pool_size =
+            application_pool_start + application_pool_size - dram_midpoint;
+        InsertPoolPartitionRegionIntoBothTrees(
+            memory_layout, application_pool_start, first_application_pool_size,
+            KMemoryRegionType_DramApplicationPool, KMemoryRegionType_VirtualDramApplicationPool,
+            cur_pool_attr);
+        InsertPoolPartitionRegionIntoBothTrees(
+            memory_layout, dram_midpoint, second_application_pool_size,
+            KMemoryRegionType_DramApplicationPool, KMemoryRegionType_VirtualDramApplicationPool,
+            cur_pool_attr);
+        total_overhead_size +=
+            KMemoryManager::CalculateManagementOverheadSize(first_application_pool_size);
+        total_overhead_size +=
+            KMemoryManager::CalculateManagementOverheadSize(second_application_pool_size);
+    }
+
+    // Insert the applet pool.
+    InsertPoolPartitionRegionIntoBothTrees(memory_layout, applet_pool_start, applet_pool_size,
+                                           KMemoryRegionType_DramAppletPool,
+                                           KMemoryRegionType_VirtualDramAppletPool, cur_pool_attr);
+    total_overhead_size += KMemoryManager::CalculateManagementOverheadSize(applet_pool_size);
+
+    // Insert the nonsecure system pool.
+    InsertPoolPartitionRegionIntoBothTrees(
+        memory_layout, unsafe_system_pool_start, unsafe_system_pool_size,
+        KMemoryRegionType_DramSystemNonSecurePool, KMemoryRegionType_VirtualDramSystemNonSecurePool,
+        cur_pool_attr);
+    total_overhead_size += KMemoryManager::CalculateManagementOverheadSize(unsafe_system_pool_size);
+
+    // Insert the pool management region.
+    total_overhead_size += KMemoryManager::CalculateManagementOverheadSize(
+        (unsafe_system_pool_start - pool_partitions_start) - total_overhead_size);
+    const u64 pool_management_start = unsafe_system_pool_start - total_overhead_size;
+    const size_t pool_management_size = total_overhead_size;
+    u32 pool_management_attr = 0;
+    InsertPoolPartitionRegionIntoBothTrees(
+        memory_layout, pool_management_start, pool_management_size,
+        KMemoryRegionType_DramPoolManagement, KMemoryRegionType_VirtualDramPoolManagement,
+        pool_management_attr);
+
+    // Insert the system pool.
+    const u64 system_pool_size = pool_management_start - pool_partitions_start;
+    InsertPoolPartitionRegionIntoBothTrees(memory_layout, pool_partitions_start, system_pool_size,
+                                           KMemoryRegionType_DramSystemPool,
+                                           KMemoryRegionType_VirtualDramSystemPool, cur_pool_attr);
+}
+
+} // namespace Init
+
+} // namespace Kernel

src/core/hle/kernel/k_memory_layout.cpp

@@ -0,0 +1,166 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <array>
+
+#include "common/alignment.h"
+#include "core/hle/kernel/k_memory_layout.h"
+#include "core/hle/kernel/k_system_control.h"
+
+namespace Kernel {
+
+namespace {
+
+template <typename... Args>
+KMemoryRegion* AllocateRegion(KMemoryRegionAllocator& memory_region_allocator, Args&&... args) {
+    return memory_region_allocator.Allocate(std::forward<Args>(args)...);
+}
+
+} // namespace
+
+KMemoryRegionTree::KMemoryRegionTree(KMemoryRegionAllocator& memory_region_allocator_)
+    : memory_region_allocator{memory_region_allocator_} {}
+
+void KMemoryRegionTree::InsertDirectly(u64 address, u64 last_address, u32 attr, u32 type_id) {
+    this->insert(*AllocateRegion(memory_region_allocator, address, last_address, attr, type_id));
+}
+
+bool KMemoryRegionTree::Insert(u64 address, size_t size, u32 type_id, u32 new_attr, u32 old_attr) {
+    // Locate the memory region that contains the address.
+    KMemoryRegion* found = this->FindModifiable(address);
+
+    // We require that the old attr is correct.
+    if (found->GetAttributes() != old_attr) {
+        return false;
+    }
+
+    // We further require that the region can be split from the old region.
+    const u64 inserted_region_end = address + size;
+    const u64 inserted_region_last = inserted_region_end - 1;
+    if (found->GetLastAddress() < inserted_region_last) {
+        return false;
+    }
+
+    // Further, we require that the type id is a valid transformation.
+    if (!found->CanDerive(type_id)) {
+        return false;
+    }
+
+    // Cache information from the region before we remove it.
+    const u64 old_address = found->GetAddress();
+    const u64 old_last = found->GetLastAddress();
+    const u64 old_pair = found->GetPairAddress();
+    const u32 old_type = found->GetType();
+
+    // Erase the existing region from the tree.
+    this->erase(this->iterator_to(*found));
+
+    // Insert the new region into the tree.
+    if (old_address == address) {
+        // Reuse the old object for the new region, if we can.
+        found->Reset(address, inserted_region_last, old_pair, new_attr, type_id);
+        this->insert(*found);
+    } else {
+        // If we can't re-use, adjust the old region.
+        found->Reset(old_address, address - 1, old_pair, old_attr, old_type);
+        this->insert(*found);
+
+        // Insert a new region for the split.
+        const u64 new_pair = (old_pair != std::numeric_limits<u64>::max())
+                                 ? old_pair + (address - old_address)
+                                 : old_pair;
+        this->insert(*AllocateRegion(memory_region_allocator, address, inserted_region_last,
+                                     new_pair, new_attr, type_id));
+    }
+
+    // If we need to insert a region after the region, do so.
+    if (old_last != inserted_region_last) {
+        const u64 after_pair = (old_pair != std::numeric_limits<u64>::max())
+                                   ? old_pair + (inserted_region_end - old_address)
+                                   : old_pair;
+        this->insert(*AllocateRegion(memory_region_allocator, inserted_region_end, old_last,
+                                     after_pair, old_attr, old_type));
+    }
+
+    return true;
+}
+
+VAddr KMemoryRegionTree::GetRandomAlignedRegion(size_t size, size_t alignment, u32 type_id) {
+    // We want to find the total extents of the type id.
+    const auto extents = this->GetDerivedRegionExtents(static_cast<KMemoryRegionType>(type_id));
+
+    // Ensure that our alignment is correct.
+    ASSERT(Common::IsAligned(extents.GetAddress(), alignment));
+
+    const u64 first_address = extents.GetAddress();
+    const u64 last_address = extents.GetLastAddress();
+
+    const u64 first_index = first_address / alignment;
+    const u64 last_index = last_address / alignment;
+
+    while (true) {
+        const u64 candidate =
+            KSystemControl::GenerateRandomRange(first_index, last_index) * alignment;
+
+        // Ensure that the candidate doesn't overflow with the size.
+        if (!(candidate < candidate + size)) {
+            continue;
+        }
+
+        const u64 candidate_last = candidate + size - 1;
+
+        // Ensure that the candidate fits within the region.
+        if (candidate_last > last_address) {
+            continue;
+        }
+
+        // Locate the candidate region, and ensure it fits and has the correct type id.
+        if (const auto& candidate_region = *this->Find(candidate);
+            !(candidate_last <= candidate_region.GetLastAddress() &&
+              candidate_region.GetType() == type_id)) {
+            continue;
+        }
+
+        return candidate;
+    }
+}
+
+KMemoryLayout::KMemoryLayout()
+    : virtual_tree{memory_region_allocator}, physical_tree{memory_region_allocator},
+      virtual_linear_tree{memory_region_allocator}, physical_linear_tree{memory_region_allocator} {}
+
+void KMemoryLayout::InitializeLinearMemoryRegionTrees(PAddr aligned_linear_phys_start,
+                                                      VAddr linear_virtual_start) {
+    // Set static differences.
+    linear_phys_to_virt_diff = linear_virtual_start - aligned_linear_phys_start;
+    linear_virt_to_phys_diff = aligned_linear_phys_start - linear_virtual_start;
+
+    // Initialize linear trees.
+    for (auto& region : GetPhysicalMemoryRegionTree()) {
+        if (region.HasTypeAttribute(KMemoryRegionAttr_LinearMapped)) {
+            GetPhysicalLinearMemoryRegionTree().InsertDirectly(
+                region.GetAddress(), region.GetLastAddress(), region.GetAttributes(),
+                region.GetType());
+        }
+    }
+
+    for (auto& region : GetVirtualMemoryRegionTree()) {
+        if (region.IsDerivedFrom(KMemoryRegionType_Dram)) {
+            GetVirtualLinearMemoryRegionTree().InsertDirectly(
+                region.GetAddress(), region.GetLastAddress(), region.GetAttributes(),
+                region.GetType());
+        }
+    }
+}
+
+size_t KMemoryLayout::GetResourceRegionSizeForInit() {
+    // Calculate resource region size based on whether we allow extra threads.
+    const bool use_extra_resources = KSystemControl::Init::ShouldIncreaseThreadResourceLimit();
+    size_t resource_region_size =
+        KernelResourceSize + (use_extra_resources ? KernelSlabHeapAdditionalSize : 0);
+
+    return resource_region_size;
+}
+
+} // namespace Kernel

src/core/hle/kernel/k_memory_layout.h

@@ -0,0 +1,399 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <utility>
+
+#include "common/alignment.h"
+#include "common/common_sizes.h"
+#include "common/common_types.h"
+#include "core/device_memory.h"
+#include "core/hle/kernel/k_memory_region.h"
+#include "core/hle/kernel/k_memory_region_type.h"
+#include "core/hle/kernel/memory_types.h"
+
+namespace Kernel {
+
+constexpr std::size_t L1BlockSize = Common::Size_1_GB;
+constexpr std::size_t L2BlockSize = Common::Size_2_MB;
+
+constexpr std::size_t GetMaximumOverheadSize(std::size_t size) {
+    return (Common::DivideUp(size, L1BlockSize) + Common::DivideUp(size, L2BlockSize)) * PageSize;
+}
+
+constexpr std::size_t MainMemorySize = Common::Size_4_GB;
+constexpr std::size_t MainMemorySizeMax = Common::Size_8_GB;
+
+constexpr std::size_t ReservedEarlyDramSize = 0x60000;
+constexpr std::size_t DramPhysicalAddress = 0x80000000;
+
+constexpr std::size_t KernelAslrAlignment = Common::Size_2_MB;
+constexpr std::size_t KernelVirtualAddressSpaceWidth = 1ULL << 39;
+constexpr std::size_t KernelPhysicalAddressSpaceWidth = 1ULL << 48;
+
+constexpr std::size_t KernelVirtualAddressSpaceBase = 0ULL - KernelVirtualAddressSpaceWidth;
+constexpr std::size_t KernelVirtualAddressSpaceEnd =
+    KernelVirtualAddressSpaceBase + (KernelVirtualAddressSpaceWidth - KernelAslrAlignment);
+constexpr std::size_t KernelVirtualAddressSpaceLast = KernelVirtualAddressSpaceEnd - 1ULL;
+constexpr std::size_t KernelVirtualAddressSpaceSize =
+    KernelVirtualAddressSpaceEnd - KernelVirtualAddressSpaceBase;
+constexpr std::size_t KernelVirtualAddressCodeBase = KernelVirtualAddressSpaceBase;
+constexpr std::size_t KernelVirtualAddressCodeSize = 0x62000;
+constexpr std::size_t KernelVirtualAddressCodeEnd =
+    KernelVirtualAddressCodeBase + KernelVirtualAddressCodeSize;
+
+constexpr std::size_t KernelPhysicalAddressSpaceBase = 0ULL;
+constexpr std::size_t KernelPhysicalAddressSpaceEnd =
+    KernelPhysicalAddressSpaceBase + KernelPhysicalAddressSpaceWidth;
+constexpr std::size_t KernelPhysicalAddressSpaceLast = KernelPhysicalAddressSpaceEnd - 1ULL;
+constexpr std::size_t KernelPhysicalAddressSpaceSize =
+    KernelPhysicalAddressSpaceEnd - KernelPhysicalAddressSpaceBase;
+constexpr std::size_t KernelPhysicalAddressCodeBase = DramPhysicalAddress + ReservedEarlyDramSize;
+
+constexpr std::size_t KernelPageTableHeapSize = GetMaximumOverheadSize(MainMemorySizeMax);
+constexpr std::size_t KernelInitialPageHeapSize = Common::Size_128_KB;
+
+constexpr std::size_t KernelSlabHeapDataSize = Common::Size_5_MB;
+constexpr std::size_t KernelSlabHeapGapsSize = Common::Size_2_MB - Common::Size_64_KB;
+constexpr std::size_t KernelSlabHeapSize = KernelSlabHeapDataSize + KernelSlabHeapGapsSize;
+
+// NOTE: This is calculated from KThread slab counts, assuming KThread size <= 0x860.
+constexpr std::size_t KernelSlabHeapAdditionalSize = 0x68000ULL;
+
+constexpr std::size_t KernelResourceSize =
+    KernelPageTableHeapSize + KernelInitialPageHeapSize + KernelSlabHeapSize;
+
+constexpr bool IsKernelAddressKey(VAddr key) {
+    return KernelVirtualAddressSpaceBase <= key && key <= KernelVirtualAddressSpaceLast;
+}
+
+constexpr bool IsKernelAddress(VAddr address) {
+    return KernelVirtualAddressSpaceBase <= address && address < KernelVirtualAddressSpaceEnd;
+}
+
+class KMemoryLayout final {
+public:
+    KMemoryLayout();
+
+    KMemoryRegionTree& GetVirtualMemoryRegionTree() {
+        return virtual_tree;
+    }
+    const KMemoryRegionTree& GetVirtualMemoryRegionTree() const {
+        return virtual_tree;
+    }
+    KMemoryRegionTree& GetPhysicalMemoryRegionTree() {
+        return physical_tree;
+    }
+    const KMemoryRegionTree& GetPhysicalMemoryRegionTree() const {
+        return physical_tree;
+    }
+    KMemoryRegionTree& GetVirtualLinearMemoryRegionTree() {
+        return virtual_linear_tree;
+    }
+    const KMemoryRegionTree& GetVirtualLinearMemoryRegionTree() const {
+        return virtual_linear_tree;
+    }
+    KMemoryRegionTree& GetPhysicalLinearMemoryRegionTree() {
+        return physical_linear_tree;
+    }
+    const KMemoryRegionTree& GetPhysicalLinearMemoryRegionTree() const {
+        return physical_linear_tree;
+    }
+
+    VAddr GetLinearVirtualAddress(PAddr address) const {
+        return address + linear_phys_to_virt_diff;
+    }
+    PAddr GetLinearPhysicalAddress(VAddr address) const {
+        return address + linear_virt_to_phys_diff;
+    }
+
+    const KMemoryRegion* FindVirtual(VAddr address) const {
+        return Find(address, GetVirtualMemoryRegionTree());
+    }
+    const KMemoryRegion* FindPhysical(PAddr address) const {
+        return Find(address, GetPhysicalMemoryRegionTree());
+    }
+
+    const KMemoryRegion* FindVirtualLinear(VAddr address) const {
+        return Find(address, GetVirtualLinearMemoryRegionTree());
+    }
+    const KMemoryRegion* FindPhysicalLinear(PAddr address) const {
+        return Find(address, GetPhysicalLinearMemoryRegionTree());
+    }
+
+    VAddr GetMainStackTopAddress(s32 core_id) const {
+        return GetStackTopAddress(core_id, KMemoryRegionType_KernelMiscMainStack);
+    }
+    VAddr GetIdleStackTopAddress(s32 core_id) const {
+        return GetStackTopAddress(core_id, KMemoryRegionType_KernelMiscIdleStack);
+    }
+    VAddr GetExceptionStackTopAddress(s32 core_id) const {
+        return GetStackTopAddress(core_id, KMemoryRegionType_KernelMiscExceptionStack);
+    }
+
+    VAddr GetSlabRegionAddress() const {
+        return Dereference(GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_KernelSlab))
+            .GetAddress();
+    }
+
+    const KMemoryRegion& GetDeviceRegion(KMemoryRegionType type) const {
+        return Dereference(GetPhysicalMemoryRegionTree().FindFirstDerived(type));
+    }
+    PAddr GetDevicePhysicalAddress(KMemoryRegionType type) const {
+        return GetDeviceRegion(type).GetAddress();
+    }
+    VAddr GetDeviceVirtualAddress(KMemoryRegionType type) const {
+        return GetDeviceRegion(type).GetPairAddress();
+    }
+
+    const KMemoryRegion& GetPoolManagementRegion() const {
+        return Dereference(
+            GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_VirtualDramPoolManagement));
+    }
+    const KMemoryRegion& GetPageTableHeapRegion() const {
+        return Dereference(
+            GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_VirtualDramKernelPtHeap));
+    }
+    const KMemoryRegion& GetKernelStackRegion() const {
+        return Dereference(GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_KernelStack));
+    }
+    const KMemoryRegion& GetTempRegion() const {
+        return Dereference(GetVirtualMemoryRegionTree().FindByType(KMemoryRegionType_KernelTemp));
+    }
+
+    const KMemoryRegion& GetKernelTraceBufferRegion() const {
+        return Dereference(GetVirtualLinearMemoryRegionTree().FindByType(
+            KMemoryRegionType_VirtualDramKernelTraceBuffer));
+    }
+
+    const KMemoryRegion& GetVirtualLinearRegion(VAddr address) const {
+        return Dereference(FindVirtualLinear(address));
+    }
+
+    const KMemoryRegion* GetPhysicalKernelTraceBufferRegion() const {
+        return GetPhysicalMemoryRegionTree().FindFirstDerived(KMemoryRegionType_KernelTraceBuffer);
+    }
+    const KMemoryRegion* GetPhysicalOnMemoryBootImageRegion() const {
+        return GetPhysicalMemoryRegionTree().FindFirstDerived(KMemoryRegionType_OnMemoryBootImage);
+    }
+    const KMemoryRegion* GetPhysicalDTBRegion() const {
+        return GetPhysicalMemoryRegionTree().FindFirstDerived(KMemoryRegionType_DTB);
+    }
+
+    bool IsHeapPhysicalAddress(const KMemoryRegion*& region, PAddr address) const {
+        return IsTypedAddress(region, address, GetPhysicalLinearMemoryRegionTree(),
+                              KMemoryRegionType_DramUserPool);
+    }
+    bool IsHeapVirtualAddress(const KMemoryRegion*& region, VAddr address) const {
+        return IsTypedAddress(region, address, GetVirtualLinearMemoryRegionTree(),
+                              KMemoryRegionType_VirtualDramUserPool);
+    }
+
+    bool IsHeapPhysicalAddress(const KMemoryRegion*& region, PAddr address, size_t size) const {
+        return IsTypedAddress(region, address, size, GetPhysicalLinearMemoryRegionTree(),
+                              KMemoryRegionType_DramUserPool);
+    }
+    bool IsHeapVirtualAddress(const KMemoryRegion*& region, VAddr address, size_t size) const {
+        return IsTypedAddress(region, address, size, GetVirtualLinearMemoryRegionTree(),
+                              KMemoryRegionType_VirtualDramUserPool);
+    }
+
+    bool IsLinearMappedPhysicalAddress(const KMemoryRegion*& region, PAddr address) const {
+        return IsTypedAddress(region, address, GetPhysicalLinearMemoryRegionTree(),
+                              static_cast<KMemoryRegionType>(KMemoryRegionAttr_LinearMapped));
+    }
+    bool IsLinearMappedPhysicalAddress(const KMemoryRegion*& region, PAddr address,
+                                       size_t size) const {
+        return IsTypedAddress(region, address, size, GetPhysicalLinearMemoryRegionTree(),
+                              static_cast<KMemoryRegionType>(KMemoryRegionAttr_LinearMapped));
+    }
+
+    std::pair<size_t, size_t> GetTotalAndKernelMemorySizes() const {
+        size_t total_size = 0, kernel_size = 0;
+        for (const auto& region : GetPhysicalMemoryRegionTree()) {
+            if (region.IsDerivedFrom(KMemoryRegionType_Dram)) {
+                total_size += region.GetSize();
+                if (!region.IsDerivedFrom(KMemoryRegionType_DramUserPool)) {
+                    kernel_size += region.GetSize();
+                }
+            }
+        }
+        return std::make_pair(total_size, kernel_size);
+    }
+
+    void InitializeLinearMemoryRegionTrees(PAddr aligned_linear_phys_start,
+                                           VAddr linear_virtual_start);
+    static size_t GetResourceRegionSizeForInit();
+
+    auto GetKernelRegionExtents() const {
+        return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_Kernel);
+    }
+    auto GetKernelCodeRegionExtents() const {
+        return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_KernelCode);
+    }
+    auto GetKernelStackRegionExtents() const {
+        return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_KernelStack);
+    }
+    auto GetKernelMiscRegionExtents() const {
+        return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_KernelMisc);
+    }
+    auto GetKernelSlabRegionExtents() const {
+        return GetVirtualMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_KernelSlab);
+    }
+
+    auto GetLinearRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionAttr_LinearMapped);
+    }
+
+    auto GetLinearRegionVirtualExtents() const {
+        const auto physical = GetLinearRegionPhysicalExtents();
+        return KMemoryRegion(GetLinearVirtualAddress(physical.GetAddress()),
+                             GetLinearVirtualAddress(physical.GetLastAddress()), 0,
+                             KMemoryRegionType_None);
+    }
+
+    auto GetMainMemoryPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(KMemoryRegionType_Dram);
+    }
+    auto GetCarveoutRegionExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionAttr_CarveoutProtected);
+    }
+
+    auto GetKernelRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramKernelBase);
+    }
+    auto GetKernelCodeRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramKernelCode);
+    }
+    auto GetKernelSlabRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramKernelSlab);
+    }
+    auto GetKernelPageTableHeapRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramKernelPtHeap);
+    }
+    auto GetKernelInitPageTableRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramKernelInitPt);
+    }
+
+    auto GetKernelPoolManagementRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramPoolManagement);
+    }
+    auto GetKernelPoolPartitionRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramPoolPartition);
+    }
+    auto GetKernelSystemPoolRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramSystemPool);
+    }
+    auto GetKernelSystemNonSecurePoolRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramSystemNonSecurePool);
+    }
+    auto GetKernelAppletPoolRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramAppletPool);
+    }
+    auto GetKernelApplicationPoolRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_DramApplicationPool);
+    }
+
+    auto GetKernelTraceBufferRegionPhysicalExtents() const {
+        return GetPhysicalMemoryRegionTree().GetDerivedRegionExtents(
+            KMemoryRegionType_KernelTraceBuffer);
+    }
+
+private:
+    template <typename AddressType>
+    static bool IsTypedAddress(const KMemoryRegion*& region, AddressType address,
+                               const KMemoryRegionTree& tree, KMemoryRegionType type) {
+        // Check if the cached region already contains the address.
+        if (region != nullptr && region->Contains(address)) {
+            return true;
+        }
+
+        // Find the containing region, and update the cache.
+        if (const KMemoryRegion* found = tree.Find(address);
+            found != nullptr && found->IsDerivedFrom(type)) {
+            region = found;
+            return true;
+        } else {
+            return false;
+        }
+    }
+
+    template <typename AddressType>
+    static bool IsTypedAddress(const KMemoryRegion*& region, AddressType address, size_t size,
+                               const KMemoryRegionTree& tree, KMemoryRegionType type) {
+        // Get the end of the checked region.
+        const u64 last_address = address + size - 1;
+
+        // Walk the tree to verify the region is correct.
+        const KMemoryRegion* cur =
+            (region != nullptr && region->Contains(address)) ? region : tree.Find(address);
+        while (cur != nullptr && cur->IsDerivedFrom(type)) {
+            if (last_address <= cur->GetLastAddress()) {
+                region = cur;
+                return true;
+            }
+
+            cur = cur->GetNext();
+        }
+        return false;
+    }
+
+    template <typename AddressType>
+    static const KMemoryRegion* Find(AddressType address, const KMemoryRegionTree& tree) {
+        return tree.Find(address);
+    }
+
+    static KMemoryRegion& Dereference(KMemoryRegion* region) {
+        ASSERT(region != nullptr);
+        return *region;
+    }
+
+    static const KMemoryRegion& Dereference(const KMemoryRegion* region) {
+        ASSERT(region != nullptr);
+        return *region;
+    }
+
+    VAddr GetStackTopAddress(s32 core_id, KMemoryRegionType type) const {
+        const auto& region = Dereference(
+            GetVirtualMemoryRegionTree().FindByTypeAndAttribute(type, static_cast<u32>(core_id)));
+        ASSERT(region.GetEndAddress() != 0);
+        return region.GetEndAddress();
+    }
+
+private:
+    u64 linear_phys_to_virt_diff{};
+    u64 linear_virt_to_phys_diff{};
+    KMemoryRegionAllocator memory_region_allocator;
+    KMemoryRegionTree virtual_tree;
+    KMemoryRegionTree physical_tree;
+    KMemoryRegionTree virtual_linear_tree;
+    KMemoryRegionTree physical_linear_tree;
+};
+
+namespace Init {
+
+// These should be generic, regardless of board.
+void SetupPoolPartitionMemoryRegions(KMemoryLayout& memory_layout);
+
+// These may be implemented in a board-specific manner.
+void SetupDevicePhysicalMemoryRegions(KMemoryLayout& memory_layout);
+void SetupDramPhysicalMemoryRegions(KMemoryLayout& memory_layout);
+
+} // namespace Init
+
+} // namespace Kernel

src/core/hle/kernel/k_memory_manager.cpp

@@ -8,20 +8,20 @@
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "common/scope_exit.h"
-#include "core/hle/kernel/memory/memory_manager.h"
-#include "core/hle/kernel/memory/page_linked_list.h"
+#include "core/hle/kernel/k_memory_manager.h"
+#include "core/hle/kernel/k_page_linked_list.h"
 #include "core/hle/kernel/svc_results.h"
 
-namespace Kernel::Memory {
+namespace Kernel {
 
-std::size_t MemoryManager::Impl::Initialize(Pool new_pool, u64 start_address, u64 end_address) {
+std::size_t KMemoryManager::Impl::Initialize(Pool new_pool, u64 start_address, u64 end_address) {
     const auto size{end_address - start_address};
 
     // Calculate metadata sizes
     const auto ref_count_size{(size / PageSize) * sizeof(u16)};
     const auto optimize_map_size{(Common::AlignUp((size / PageSize), 64) / 64) * sizeof(u64)};
     const auto manager_size{Common::AlignUp(optimize_map_size + ref_count_size, PageSize)};
-    const auto page_heap_size{PageHeap::CalculateMetadataOverheadSize(size)};
+    const auto page_heap_size{KPageHeap::CalculateManagementOverheadSize(size)};
     const auto total_metadata_size{manager_size + page_heap_size};
     ASSERT(manager_size <= total_metadata_size);
     ASSERT(Common::IsAligned(total_metadata_size, PageSize));
@@ -41,29 +41,30 @@ std::size_t MemoryManager::Impl::Initialize(Pool new_pool, u64 start_address, u6
     return total_metadata_size;
 }
 
-void MemoryManager::InitializeManager(Pool pool, u64 start_address, u64 end_address) {
+void KMemoryManager::InitializeManager(Pool pool, u64 start_address, u64 end_address) {
     ASSERT(pool < Pool::Count);
     managers[static_cast<std::size_t>(pool)].Initialize(pool, start_address, end_address);
 }
 
-VAddr MemoryManager::AllocateContinuous(std::size_t num_pages, std::size_t align_pages, Pool pool,
-                                        Direction dir) {
+VAddr KMemoryManager::AllocateAndOpenContinuous(std::size_t num_pages, std::size_t align_pages,
+                                                u32 option) {
     // Early return if we're allocating no pages
     if (num_pages == 0) {
         return {};
     }
 
     // Lock the pool that we're allocating from
+    const auto [pool, dir] = DecodeOption(option);
     const auto pool_index{static_cast<std::size_t>(pool)};
     std::lock_guard lock{pool_locks[pool_index]};
 
     // Choose a heap based on our page size request
-    const s32 heap_index{PageHeap::GetAlignedBlockIndex(num_pages, align_pages)};
+    const s32 heap_index{KPageHeap::GetAlignedBlockIndex(num_pages, align_pages)};
 
     // Loop, trying to iterate from each block
     // TODO (bunnei): Support multiple managers
     Impl& chosen_manager{managers[pool_index]};
-    VAddr allocated_block{chosen_manager.AllocateBlock(heap_index)};
+    VAddr allocated_block{chosen_manager.AllocateBlock(heap_index, false)};
 
     // If we failed to allocate, quit now
     if (!allocated_block) {
@@ -71,7 +72,7 @@ VAddr MemoryManager::AllocateContinuous(std::size_t num_pages, std::size_t align
     }
 
     // If we allocated more than we need, free some
-    const auto allocated_pages{PageHeap::GetBlockNumPages(heap_index)};
+    const auto allocated_pages{KPageHeap::GetBlockNumPages(heap_index)};
     if (allocated_pages > num_pages) {
         chosen_manager.Free(allocated_block + num_pages * PageSize, allocated_pages - num_pages);
     }
@@ -79,8 +80,8 @@ VAddr MemoryManager::AllocateContinuous(std::size_t num_pages, std::size_t align
     return allocated_block;
 }
 
-ResultCode MemoryManager::Allocate(PageLinkedList& page_list, std::size_t num_pages, Pool pool,
-                                   Direction dir) {
+ResultCode KMemoryManager::Allocate(KPageLinkedList& page_list, std::size_t num_pages, Pool pool,
+                                    Direction dir) {
     ASSERT(page_list.GetNumPages() == 0);
 
     // Early return if we're allocating no pages
@@ -93,7 +94,7 @@ ResultCode MemoryManager::Allocate(PageLinkedList& page_list, std::size_t num_pa
     std::lock_guard lock{pool_locks[pool_index]};
 
     // Choose a heap based on our page size request
-    const s32 heap_index{PageHeap::GetBlockIndex(num_pages)};
+    const s32 heap_index{KPageHeap::GetBlockIndex(num_pages)};
     if (heap_index < 0) {
         return ResultOutOfMemory;
     }
@@ -112,11 +113,11 @@ ResultCode MemoryManager::Allocate(PageLinkedList& page_list, std::size_t num_pa
 
     // Keep allocating until we've allocated all our pages
     for (s32 index{heap_index}; index >= 0 && num_pages > 0; index--) {
-        const auto pages_per_alloc{PageHeap::GetBlockNumPages(index)};
+        const auto pages_per_alloc{KPageHeap::GetBlockNumPages(index)};
 
         while (num_pages >= pages_per_alloc) {
             // Allocate a block
-            VAddr allocated_block{chosen_manager.AllocateBlock(index)};
+            VAddr allocated_block{chosen_manager.AllocateBlock(index, false)};
             if (!allocated_block) {
                 break;
             }
@@ -148,8 +149,8 @@ ResultCode MemoryManager::Allocate(PageLinkedList& page_list, std::size_t num_pa
     return RESULT_SUCCESS;
 }
 
-ResultCode MemoryManager::Free(PageLinkedList& page_list, std::size_t num_pages, Pool pool,
-                               Direction dir) {
+ResultCode KMemoryManager::Free(KPageLinkedList& page_list, std::size_t num_pages, Pool pool,
+                                Direction dir) {
     // Early return if we're freeing no pages
     if (!num_pages) {
         return RESULT_SUCCESS;
@@ -172,4 +173,16 @@ ResultCode MemoryManager::Free(PageLinkedList& page_list, std::size_t num_pages,
     return RESULT_SUCCESS;
 }
 
-} // namespace Kernel::Memory
+std::size_t KMemoryManager::Impl::CalculateManagementOverheadSize(std::size_t region_size) {
+    const std::size_t ref_count_size = (region_size / PageSize) * sizeof(u16);
+    const std::size_t optimize_map_size =
+        (Common::AlignUp((region_size / PageSize), Common::BitSize<u64>()) /
+         Common::BitSize<u64>()) *
+        sizeof(u64);
+    const std::size_t manager_meta_size =
+        Common::AlignUp(optimize_map_size + ref_count_size, PageSize);
+    const std::size_t page_heap_size = KPageHeap::CalculateManagementOverheadSize(region_size);
+    return manager_meta_size + page_heap_size;
+}
+
+} // namespace Kernel

src/core/hle/kernel/k_memory_manager.h

@@ -0,0 +1,137 @@
+// Copyright 2020 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <mutex>
+#include <tuple>
+
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+#include "core/hle/kernel/k_page_heap.h"
+#include "core/hle/result.h"
+
+namespace Kernel {
+
+class KPageLinkedList;
+
+class KMemoryManager final : NonCopyable {
+public:
+    enum class Pool : u32 {
+        Application = 0,
+        Applet = 1,
+        System = 2,
+        SystemNonSecure = 3,
+
+        Count,
+
+        Shift = 4,
+        Mask = (0xF << Shift),
+
+        // Aliases.
+        Unsafe = Application,
+        Secure = System,
+    };
+
+    enum class Direction : u32 {
+        FromFront = 0,
+        FromBack = 1,
+
+        Shift = 0,
+        Mask = (0xF << Shift),
+    };
+
+    KMemoryManager() = default;
+
+    constexpr std::size_t GetSize(Pool pool) const {
+        return managers[static_cast<std::size_t>(pool)].GetSize();
+    }
+
+    void InitializeManager(Pool pool, u64 start_address, u64 end_address);
+
+    VAddr AllocateAndOpenContinuous(size_t num_pages, size_t align_pages, u32 option);
+    ResultCode Allocate(KPageLinkedList& page_list, std::size_t num_pages, Pool pool,
+                        Direction dir = Direction::FromFront);
+    ResultCode Free(KPageLinkedList& page_list, std::size_t num_pages, Pool pool,
+                    Direction dir = Direction::FromFront);
+
+    static constexpr std::size_t MaxManagerCount = 10;
+
+public:
+    static std::size_t CalculateManagementOverheadSize(std::size_t region_size) {
+        return Impl::CalculateManagementOverheadSize(region_size);
+    }
+
+    static constexpr u32 EncodeOption(Pool pool, Direction dir) {
+        return (static_cast<u32>(pool) << static_cast<u32>(Pool::Shift)) |
+               (static_cast<u32>(dir) << static_cast<u32>(Direction::Shift));
+    }
+
+    static constexpr Pool GetPool(u32 option) {
+        return static_cast<Pool>((static_cast<u32>(option) & static_cast<u32>(Pool::Mask)) >>
+                                 static_cast<u32>(Pool::Shift));
+    }
+
+    static constexpr Direction GetDirection(u32 option) {
+        return static_cast<Direction>(
+            (static_cast<u32>(option) & static_cast<u32>(Direction::Mask)) >>
+            static_cast<u32>(Direction::Shift));
+    }
+
+    static constexpr std::tuple<Pool, Direction> DecodeOption(u32 option) {
+        return std::make_tuple(GetPool(option), GetDirection(option));
+    }
+
+private:
+    class Impl final : NonCopyable {
+    private:
+        using RefCount = u16;
+
+    private:
+        KPageHeap heap;
+        Pool pool{};
+
+    public:
+        static std::size_t CalculateManagementOverheadSize(std::size_t region_size);
+
+        static constexpr std::size_t CalculateOptimizedProcessOverheadSize(
+            std::size_t region_size) {
+            return (Common::AlignUp((region_size / PageSize), Common::BitSize<u64>()) /
+                    Common::BitSize<u64>()) *
+                   sizeof(u64);
+        }
+
+    public:
+        Impl() = default;
+
+        std::size_t Initialize(Pool new_pool, u64 start_address, u64 end_address);
+
+        VAddr AllocateBlock(s32 index, bool random) {
+            return heap.AllocateBlock(index, random);
+        }
+
+        void Free(VAddr addr, std::size_t num_pages) {
+            heap.Free(addr, num_pages);
+        }
+
+        constexpr std::size_t GetSize() const {
+            return heap.GetSize();
+        }
+
+        constexpr VAddr GetAddress() const {
+            return heap.GetAddress();
+        }
+
+        constexpr VAddr GetEndAddress() const {
+            return heap.GetEndAddress();
+        }
+    };
+
+private:
+    std::array<std::mutex, static_cast<std::size_t>(Pool::Count)> pool_locks;
+    std::array<Impl, MaxManagerCount> managers;
+};
+
+} // namespace Kernel

src/core/hle/kernel/k_memory_region.h

@@ -0,0 +1,350 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/assert.h"
+#include "common/common_types.h"
+#include "common/intrusive_red_black_tree.h"
+#include "core/hle/kernel/k_memory_region_type.h"
+
+namespace Kernel {
+
+class KMemoryRegionAllocator;
+
+class KMemoryRegion final : public Common::IntrusiveRedBlackTreeBaseNode<KMemoryRegion>,
+                            NonCopyable {
+    friend class KMemoryRegionTree;
+
+public:
+    constexpr KMemoryRegion() = default;
+    constexpr KMemoryRegion(u64 address_, u64 last_address_)
+        : address{address_}, last_address{last_address_} {}
+    constexpr KMemoryRegion(u64 address_, u64 last_address_, u64 pair_address_, u32 attributes_,
+                            u32 type_id_)
+        : address(address_), last_address(last_address_), pair_address(pair_address_),
+          attributes(attributes_), type_id(type_id_) {}
+    constexpr KMemoryRegion(u64 address_, u64 last_address_, u32 attributes_, u32 type_id_)
+        : KMemoryRegion(address_, last_address_, std::numeric_limits<u64>::max(), attributes_,
+                        type_id_) {}
+
+    static constexpr int Compare(const KMemoryRegion& lhs, const KMemoryRegion& rhs) {
+        if (lhs.GetAddress() < rhs.GetAddress()) {
+            return -1;
+        } else if (lhs.GetAddress() <= rhs.GetLastAddress()) {
+            return 0;
+        } else {
+            return 1;
+        }
+    }
+
+private:
+    constexpr void Reset(u64 a, u64 la, u64 p, u32 r, u32 t) {
+        address = a;
+        pair_address = p;
+        last_address = la;
+        attributes = r;
+        type_id = t;
+    }
+
+public:
+    constexpr u64 GetAddress() const {
+        return address;
+    }
+
+    constexpr u64 GetPairAddress() const {
+        return pair_address;
+    }
+
+    constexpr u64 GetLastAddress() const {
+        return last_address;
+    }
+
+    constexpr u64 GetEndAddress() const {
+        return this->GetLastAddress() + 1;
+    }
+
+    constexpr size_t GetSize() const {
+        return this->GetEndAddress() - this->GetAddress();
+    }
+
+    constexpr u32 GetAttributes() const {
+        return attributes;
+    }
+
+    constexpr u32 GetType() const {
+        return type_id;
+    }
+
+    constexpr void SetType(u32 type) {
+        ASSERT(this->CanDerive(type));
+        type_id = type;
+    }
+
+    constexpr bool Contains(u64 address) const {
+        ASSERT(this->GetEndAddress() != 0);
+        return this->GetAddress() <= address && address <= this->GetLastAddress();
+    }
+
+    constexpr bool IsDerivedFrom(u32 type) const {
+        return (this->GetType() | type) == this->GetType();
+    }
+
+    constexpr bool HasTypeAttribute(u32 attr) const {
+        return (this->GetType() | attr) == this->GetType();
+    }
+
+    constexpr bool CanDerive(u32 type) const {
+        return (this->GetType() | type) == type;
+    }
+
+    constexpr void SetPairAddress(u64 a) {
+        pair_address = a;
+    }
+
+    constexpr void SetTypeAttribute(u32 attr) {
+        type_id |= attr;
+    }
+
+private:
+    u64 address{};
+    u64 last_address{};
+    u64 pair_address{};
+    u32 attributes{};
+    u32 type_id{};
+};
+
+class KMemoryRegionTree final : NonCopyable {
+public:
+    struct DerivedRegionExtents {
+        const KMemoryRegion* first_region{};
+        const KMemoryRegion* last_region{};
+
+        constexpr DerivedRegionExtents() = default;
+
+        constexpr u64 GetAddress() const {
+            return this->first_region->GetAddress();
+        }
+
+        constexpr u64 GetLastAddress() const {
+            return this->last_region->GetLastAddress();
+        }
+
+        constexpr u64 GetEndAddress() const {
+            return this->GetLastAddress() + 1;
+        }
+
+        constexpr size_t GetSize() const {
+            return this->GetEndAddress() - this->GetAddress();
+        }
+    };
+
+private:
+    using TreeType =
+        Common::IntrusiveRedBlackTreeBaseTraits<KMemoryRegion>::TreeType<KMemoryRegion>;
+
+public:
+    using value_type = TreeType::value_type;
+    using size_type = TreeType::size_type;
+    using difference_type = TreeType::difference_type;
+    using pointer = TreeType::pointer;
+    using const_pointer = TreeType::const_pointer;
+    using reference = TreeType::reference;
+    using const_reference = TreeType::const_reference;
+    using iterator = TreeType::iterator;
+    using const_iterator = TreeType::const_iterator;
+
+private:
+    TreeType m_tree{};
+    KMemoryRegionAllocator& memory_region_allocator;
+
+public:
+    explicit KMemoryRegionTree(KMemoryRegionAllocator& memory_region_allocator_);
+
+public:
+    KMemoryRegion* FindModifiable(u64 address) {
+        if (auto it = this->find(KMemoryRegion(address, address, 0, 0)); it != this->end()) {
+            return std::addressof(*it);
+        } else {
+            return nullptr;
+        }
+    }
+
+    const KMemoryRegion* Find(u64 address) const {
+        if (auto it = this->find(KMemoryRegion(address, address, 0, 0)); it != this->cend()) {
+            return std::addressof(*it);
+        } else {
+            return nullptr;
+        }
+    }
+
+    const KMemoryRegion* FindByType(KMemoryRegionType type_id) const {
+        for (auto it = this->cbegin(); it != this->cend(); ++it) {
+            if (it->GetType() == static_cast<u32>(type_id)) {
+                return std::addressof(*it);
+            }
+        }
+        return nullptr;
+    }
+
+    const KMemoryRegion* FindByTypeAndAttribute(u32 type_id, u32 attr) const {
+        for (auto it = this->cbegin(); it != this->cend(); ++it) {
+            if (it->GetType() == type_id && it->GetAttributes() == attr) {
+                return std::addressof(*it);
+            }
+        }
+        return nullptr;
+    }
+
+    const KMemoryRegion* FindFirstDerived(KMemoryRegionType type_id) const {
+        for (auto it = this->cbegin(); it != this->cend(); it++) {
+            if (it->IsDerivedFrom(type_id)) {
+                return std::addressof(*it);
+            }
+        }
+        return nullptr;
+    }
+
+    const KMemoryRegion* FindLastDerived(KMemoryRegionType type_id) const {
+        const KMemoryRegion* region = nullptr;
+        for (auto it = this->begin(); it != this->end(); it++) {
+            if (it->IsDerivedFrom(type_id)) {
+                region = std::addressof(*it);
+            }
+        }
+        return region;
+    }
+
+    DerivedRegionExtents GetDerivedRegionExtents(KMemoryRegionType type_id) const {
+        DerivedRegionExtents extents;
+
+        ASSERT(extents.first_region == nullptr);
+        ASSERT(extents.last_region == nullptr);
+
+        for (auto it = this->cbegin(); it != this->cend(); it++) {
+            if (it->IsDerivedFrom(type_id)) {
+                if (extents.first_region == nullptr) {
+                    extents.first_region = std::addressof(*it);
+                }
+                extents.last_region = std::addressof(*it);
+            }
+        }
+
+        ASSERT(extents.first_region != nullptr);
+        ASSERT(extents.last_region != nullptr);
+
+        return extents;
+    }
+
+    DerivedRegionExtents GetDerivedRegionExtents(u32 type_id) const {
+        return GetDerivedRegionExtents(static_cast<KMemoryRegionType>(type_id));
+    }
+
+public:
+    void InsertDirectly(u64 address, u64 last_address, u32 attr = 0, u32 type_id = 0);
+    bool Insert(u64 address, size_t size, u32 type_id, u32 new_attr = 0, u32 old_attr = 0);
+
+    VAddr GetRandomAlignedRegion(size_t size, size_t alignment, u32 type_id);
+
+    VAddr GetRandomAlignedRegionWithGuard(size_t size, size_t alignment, u32 type_id,
+                                          size_t guard_size) {
+        return this->GetRandomAlignedRegion(size + 2 * guard_size, alignment, type_id) + guard_size;
+    }
+
+public:
+    // Iterator accessors.
+    iterator begin() {
+        return m_tree.begin();
+    }
+
+    const_iterator begin() const {
+        return m_tree.begin();
+    }
+
+    iterator end() {
+        return m_tree.end();
+    }
+
+    const_iterator end() const {
+        return m_tree.end();
+    }
+
+    const_iterator cbegin() const {
+        return this->begin();
+    }
+
+    const_iterator cend() const {
+        return this->end();
+    }
+
+    iterator iterator_to(reference ref) {
+        return m_tree.iterator_to(ref);
+    }
+
+    const_iterator iterator_to(const_reference ref) const {
+        return m_tree.iterator_to(ref);
+    }
+
+    // Content management.
+    bool empty() const {
+        return m_tree.empty();
+    }
+
+    reference back() {
+        return m_tree.back();
+    }
+
+    const_reference back() const {
+        return m_tree.back();
+    }
+
+    reference front() {
+        return m_tree.front();
+    }
+
+    const_reference front() const {
+        return m_tree.front();
+    }
+
+    iterator insert(reference ref) {
+        return m_tree.insert(ref);
+    }
+
+    iterator erase(iterator it) {
+        return m_tree.erase(it);
+    }
+
+    iterator find(const_reference ref) const {
+        return m_tree.find(ref);
+    }
+
+    iterator nfind(const_reference ref) const {
+        return m_tree.nfind(ref);
+    }
+};
+
+class KMemoryRegionAllocator final : NonCopyable {
+public:
+    static constexpr size_t MaxMemoryRegions = 200;
+
+    constexpr KMemoryRegionAllocator() = default;
+
+    template <typename... Args>
+    KMemoryRegion* Allocate(Args&&... args) {
+        // Ensure we stay within the bounds of our heap.
+        ASSERT(this->num_regions < MaxMemoryRegions);
+
+        // Create the new region.
+        KMemoryRegion* region = std::addressof(this->region_heap[this->num_regions++]);
+        new (region) KMemoryRegion(std::forward<Args>(args)...);
+
+        return region;
+    }
+
+private:
+    std::array<KMemoryRegion, MaxMemoryRegions> region_heap{};
+    size_t num_regions{};
+};
+
+} // namespace Kernel

src/core/hle/kernel/k_memory_region_type.h

@@ -0,0 +1,338 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/bit_util.h"
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+
+#define ARCH_ARM64
+#define BOARD_NINTENDO_NX
+
+namespace Kernel {
+
+enum KMemoryRegionType : u32 {
+    KMemoryRegionAttr_CarveoutProtected = 0x04000000,
+    KMemoryRegionAttr_DidKernelMap = 0x08000000,
+    KMemoryRegionAttr_ShouldKernelMap = 0x10000000,
+    KMemoryRegionAttr_UserReadOnly = 0x20000000,
+    KMemoryRegionAttr_NoUserMap = 0x40000000,
+    KMemoryRegionAttr_LinearMapped = 0x80000000,
+};
+DECLARE_ENUM_FLAG_OPERATORS(KMemoryRegionType);
+
+namespace impl {
+
+constexpr size_t BitsForDeriveSparse(size_t n) {
+    return n + 1;
+}
+
+constexpr size_t BitsForDeriveDense(size_t n) {
+    size_t low = 0, high = 1;
+    for (size_t i = 0; i < n - 1; ++i) {
+        if ((++low) == high) {
+            ++high;
+            low = 0;
+        }
+    }
+    return high + 1;
+}
+
+class KMemoryRegionTypeValue {
+public:
+    using ValueType = std::underlying_type_t<KMemoryRegionType>;
+
+    constexpr KMemoryRegionTypeValue() = default;
+
+    constexpr operator KMemoryRegionType() const {
+        return static_cast<KMemoryRegionType>(m_value);
+    }
+
+    constexpr ValueType GetValue() const {
+        return m_value;
+    }
+
+    constexpr const KMemoryRegionTypeValue& Finalize() {
+        m_finalized = true;
+        return *this;
+    }
+
+    constexpr const KMemoryRegionTypeValue& SetSparseOnly() {
+        m_sparse_only = true;
+        return *this;
+    }
+
+    constexpr const KMemoryRegionTypeValue& SetDenseOnly() {
+        m_dense_only = true;
+        return *this;
+    }
+
+    constexpr KMemoryRegionTypeValue& SetAttribute(u32 attr) {
+        m_value |= attr;
+        return *this;
+    }
+
+    constexpr KMemoryRegionTypeValue DeriveInitial(
+        size_t i, size_t next = Common::BitSize<ValueType>()) const {
+        KMemoryRegionTypeValue new_type = *this;
+        new_type.m_value = (ValueType{1} << i);
+        new_type.m_next_bit = next;
+        return new_type;
+    }
+
+    constexpr KMemoryRegionTypeValue DeriveAttribute(u32 attr) const {
+        KMemoryRegionTypeValue new_type = *this;
+        new_type.m_value |= attr;
+        return new_type;
+    }
+
+    constexpr KMemoryRegionTypeValue DeriveTransition(size_t ofs = 0, size_t adv = 1) const {
+        KMemoryRegionTypeValue new_type = *this;
+        new_type.m_value |= (ValueType{1} << (m_next_bit + ofs));
+        new_type.m_next_bit += adv;
+        return new_type;
+    }
+
+    constexpr KMemoryRegionTypeValue DeriveSparse(size_t ofs, size_t n, size_t i) const {
+        KMemoryRegionTypeValue new_type = *this;
+        new_type.m_value |= (ValueType{1} << (m_next_bit + ofs));
+        new_type.m_value |= (ValueType{1} << (m_next_bit + ofs + 1 + i));
+        new_type.m_next_bit += ofs + n + 1;
+        return new_type;
+    }
+
+    constexpr KMemoryRegionTypeValue Derive(size_t n, size_t i) const {
+        size_t low = 0, high = 1;
+        for (size_t j = 0; j < i; ++j) {
+            if ((++low) == high) {
+                ++high;
+                low = 0;
+            }
+        }
+
+        KMemoryRegionTypeValue new_type = *this;
+        new_type.m_value |= (ValueType{1} << (m_next_bit + low));
+        new_type.m_value |= (ValueType{1} << (m_next_bit + high));
+        new_type.m_next_bit += BitsForDeriveDense(n);
+        return new_type;
+    }
+
+    constexpr KMemoryRegionTypeValue Advance(size_t n) const {
+        KMemoryRegionTypeValue new_type = *this;
+        new_type.m_next_bit += n;
+        return new_type;
+    }
+
+    constexpr bool IsAncestorOf(ValueType v) const {
+        return (m_value | v) == v;
+    }
+
+private:
+    constexpr KMemoryRegionTypeValue(ValueType v) : m_value(v) {}
+
+private:
+    ValueType m_value{};
+    size_t m_next_bit{};
+    bool m_finalized{};
+    bool m_sparse_only{};
+    bool m_dense_only{};
+};
+
+} // namespace impl
+
+constexpr auto KMemoryRegionType_None = impl::KMemoryRegionTypeValue();
+constexpr auto KMemoryRegionType_Kernel = KMemoryRegionType_None.DeriveInitial(0, 2);
+constexpr auto KMemoryRegionType_Dram = KMemoryRegionType_None.DeriveInitial(1, 2);
+static_assert(KMemoryRegionType_Kernel.GetValue() == 0x1);
+static_assert(KMemoryRegionType_Dram.GetValue() == 0x2);
+
+constexpr auto KMemoryRegionType_DramKernelBase =
+    KMemoryRegionType_Dram.DeriveSparse(0, 3, 0)
+        .SetAttribute(KMemoryRegionAttr_NoUserMap)
+        .SetAttribute(KMemoryRegionAttr_CarveoutProtected);
+constexpr auto KMemoryRegionType_DramReservedBase = KMemoryRegionType_Dram.DeriveSparse(0, 3, 1);
+constexpr auto KMemoryRegionType_DramHeapBase =
+    KMemoryRegionType_Dram.DeriveSparse(0, 3, 2).SetAttribute(KMemoryRegionAttr_LinearMapped);
+static_assert(KMemoryRegionType_DramKernelBase.GetValue() ==
+              (0xE | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_DramReservedBase.GetValue() == (0x16));
+static_assert(KMemoryRegionType_DramHeapBase.GetValue() == (0x26 | KMemoryRegionAttr_LinearMapped));
+
+constexpr auto KMemoryRegionType_DramKernelCode =
+    KMemoryRegionType_DramKernelBase.DeriveSparse(0, 4, 0);
+constexpr auto KMemoryRegionType_DramKernelSlab =
+    KMemoryRegionType_DramKernelBase.DeriveSparse(0, 4, 1);
+constexpr auto KMemoryRegionType_DramKernelPtHeap =
+    KMemoryRegionType_DramKernelBase.DeriveSparse(0, 4, 2).SetAttribute(
+        KMemoryRegionAttr_LinearMapped);
+constexpr auto KMemoryRegionType_DramKernelInitPt =
+    KMemoryRegionType_DramKernelBase.DeriveSparse(0, 4, 3).SetAttribute(
+        KMemoryRegionAttr_LinearMapped);
+static_assert(KMemoryRegionType_DramKernelCode.GetValue() ==
+              (0xCE | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_DramKernelSlab.GetValue() ==
+              (0x14E | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_DramKernelPtHeap.GetValue() ==
+              (0x24E | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap |
+               KMemoryRegionAttr_LinearMapped));
+static_assert(KMemoryRegionType_DramKernelInitPt.GetValue() ==
+              (0x44E | KMemoryRegionAttr_CarveoutProtected | KMemoryRegionAttr_NoUserMap |
+               KMemoryRegionAttr_LinearMapped));
+
+constexpr auto KMemoryRegionType_DramReservedEarly =
+    KMemoryRegionType_DramReservedBase.DeriveAttribute(KMemoryRegionAttr_NoUserMap);
+static_assert(KMemoryRegionType_DramReservedEarly.GetValue() ==
+              (0x16 | KMemoryRegionAttr_NoUserMap));
+
+constexpr auto KMemoryRegionType_KernelTraceBuffer =
+    KMemoryRegionType_DramReservedBase.DeriveSparse(0, 3, 0)
+        .SetAttribute(KMemoryRegionAttr_LinearMapped)
+        .SetAttribute(KMemoryRegionAttr_UserReadOnly);
+constexpr auto KMemoryRegionType_OnMemoryBootImage =
+    KMemoryRegionType_DramReservedBase.DeriveSparse(0, 3, 1);
+constexpr auto KMemoryRegionType_DTB = KMemoryRegionType_DramReservedBase.DeriveSparse(0, 3, 2);
+static_assert(KMemoryRegionType_KernelTraceBuffer.GetValue() ==
+              (0xD6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_UserReadOnly));
+static_assert(KMemoryRegionType_OnMemoryBootImage.GetValue() == 0x156);
+static_assert(KMemoryRegionType_DTB.GetValue() == 0x256);
+
+constexpr auto KMemoryRegionType_DramPoolPartition =
+    KMemoryRegionType_DramHeapBase.DeriveAttribute(KMemoryRegionAttr_NoUserMap);
+static_assert(KMemoryRegionType_DramPoolPartition.GetValue() ==
+              (0x26 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap));
+
+constexpr auto KMemoryRegionType_DramPoolManagement =
+    KMemoryRegionType_DramPoolPartition.DeriveTransition(0, 2).DeriveTransition().SetAttribute(
+        KMemoryRegionAttr_CarveoutProtected);
+constexpr auto KMemoryRegionType_DramUserPool =
+    KMemoryRegionType_DramPoolPartition.DeriveTransition(1, 2).DeriveTransition();
+static_assert(KMemoryRegionType_DramPoolManagement.GetValue() ==
+              (0x166 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap |
+               KMemoryRegionAttr_CarveoutProtected));
+static_assert(KMemoryRegionType_DramUserPool.GetValue() ==
+              (0x1A6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap));
+
+constexpr auto KMemoryRegionType_DramApplicationPool = KMemoryRegionType_DramUserPool.Derive(4, 0);
+constexpr auto KMemoryRegionType_DramAppletPool = KMemoryRegionType_DramUserPool.Derive(4, 1);
+constexpr auto KMemoryRegionType_DramSystemNonSecurePool =
+    KMemoryRegionType_DramUserPool.Derive(4, 2);
+constexpr auto KMemoryRegionType_DramSystemPool =
+    KMemoryRegionType_DramUserPool.Derive(4, 3).SetAttribute(KMemoryRegionAttr_CarveoutProtected);
+static_assert(KMemoryRegionType_DramApplicationPool.GetValue() ==
+              (0x7A6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_DramAppletPool.GetValue() ==
+              (0xBA6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_DramSystemNonSecurePool.GetValue() ==
+              (0xDA6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap));
+static_assert(KMemoryRegionType_DramSystemPool.GetValue() ==
+              (0x13A6 | KMemoryRegionAttr_LinearMapped | KMemoryRegionAttr_NoUserMap |
+               KMemoryRegionAttr_CarveoutProtected));
+
+constexpr auto KMemoryRegionType_VirtualDramHeapBase = KMemoryRegionType_Dram.DeriveSparse(1, 3, 0);
+constexpr auto KMemoryRegionType_VirtualDramKernelPtHeap =
+    KMemoryRegionType_Dram.DeriveSparse(1, 3, 1);
+constexpr auto KMemoryRegionType_VirtualDramKernelTraceBuffer =
+    KMemoryRegionType_Dram.DeriveSparse(1, 3, 2);
+static_assert(KMemoryRegionType_VirtualDramHeapBase.GetValue() == 0x1A);
+static_assert(KMemoryRegionType_VirtualDramKernelPtHeap.GetValue() == 0x2A);
+static_assert(KMemoryRegionType_VirtualDramKernelTraceBuffer.GetValue() == 0x4A);
+
+constexpr auto KMemoryRegionType_VirtualDramKernelInitPt =
+    KMemoryRegionType_VirtualDramHeapBase.Derive(3, 0);
+constexpr auto KMemoryRegionType_VirtualDramPoolManagement =
+    KMemoryRegionType_VirtualDramHeapBase.Derive(3, 1);
+constexpr auto KMemoryRegionType_VirtualDramUserPool =
+    KMemoryRegionType_VirtualDramHeapBase.Derive(3, 2);
+static_assert(KMemoryRegionType_VirtualDramKernelInitPt.GetValue() == 0x19A);
+static_assert(KMemoryRegionType_VirtualDramPoolManagement.GetValue() == 0x29A);
+static_assert(KMemoryRegionType_VirtualDramUserPool.GetValue() == 0x31A);
+
+// NOTE: For unknown reason, the pools are derived out-of-order here. It's worth eventually trying
+// to understand why Nintendo made this choice.
+// UNUSED: .Derive(6, 0);
+// UNUSED: .Derive(6, 1);
+constexpr auto KMemoryRegionType_VirtualDramAppletPool =
+    KMemoryRegionType_VirtualDramUserPool.Derive(6, 2);
+constexpr auto KMemoryRegionType_VirtualDramApplicationPool =
+    KMemoryRegionType_VirtualDramUserPool.Derive(6, 3);
+constexpr auto KMemoryRegionType_VirtualDramSystemNonSecurePool =
+    KMemoryRegionType_VirtualDramUserPool.Derive(6, 4);
+constexpr auto KMemoryRegionType_VirtualDramSystemPool =
+    KMemoryRegionType_VirtualDramUserPool.Derive(6, 5);
+static_assert(KMemoryRegionType_VirtualDramAppletPool.GetValue() == 0x1B1A);
+static_assert(KMemoryRegionType_VirtualDramApplicationPool.GetValue() == 0x271A);
+static_assert(KMemoryRegionType_VirtualDramSystemNonSecurePool.GetValue() == 0x2B1A);
+static_assert(KMemoryRegionType_VirtualDramSystemPool.GetValue() == 0x331A);
+
+constexpr auto KMemoryRegionType_ArchDeviceBase =
+    KMemoryRegionType_Kernel.DeriveTransition(0, 1).SetSparseOnly();
+constexpr auto KMemoryRegionType_BoardDeviceBase =
+    KMemoryRegionType_Kernel.DeriveTransition(0, 2).SetDenseOnly();
+static_assert(KMemoryRegionType_ArchDeviceBase.GetValue() == 0x5);
+static_assert(KMemoryRegionType_BoardDeviceBase.GetValue() == 0x5);
+
+#if defined(ARCH_ARM64)
+#include "core/hle/kernel/arch/arm64/k_memory_region_device_types.inc"
+#elif defined(ARCH_ARM)
+#error "Unimplemented"
+#else
+// Default to no architecture devices.
+constexpr auto NumArchitectureDeviceRegions = 0;
+#endif
+static_assert(NumArchitectureDeviceRegions >= 0);
+
+#if defined(BOARD_NINTENDO_NX)
+#include "core/hle/kernel/board/nintendo/nx/k_memory_region_device_types.inc"
+#else
+// Default to no board devices.
+constexpr auto NumBoardDeviceRegions = 0;
+#endif
+static_assert(NumBoardDeviceRegions >= 0);
+
+constexpr auto KMemoryRegionType_KernelCode = KMemoryRegionType_Kernel.DeriveSparse(1, 4, 0);
+constexpr auto KMemoryRegionType_KernelStack = KMemoryRegionType_Kernel.DeriveSparse(1, 4, 1);
+constexpr auto KMemoryRegionType_KernelMisc = KMemoryRegionType_Kernel.DeriveSparse(1, 4, 2);
+constexpr auto KMemoryRegionType_KernelSlab = KMemoryRegionType_Kernel.DeriveSparse(1, 4, 3);
+static_assert(KMemoryRegionType_KernelCode.GetValue() == 0x19);
+static_assert(KMemoryRegionType_KernelStack.GetValue() == 0x29);
+static_assert(KMemoryRegionType_KernelMisc.GetValue() == 0x49);
+static_assert(KMemoryRegionType_KernelSlab.GetValue() == 0x89);
+
+constexpr auto KMemoryRegionType_KernelMiscDerivedBase =
+    KMemoryRegionType_KernelMisc.DeriveTransition();
+static_assert(KMemoryRegionType_KernelMiscDerivedBase.GetValue() == 0x149);
+
+// UNUSED: .Derive(7, 0);
+constexpr auto KMemoryRegionType_KernelMiscMainStack =
+    KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 1);
+constexpr auto KMemoryRegionType_KernelMiscMappedDevice =
+    KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 2);
+constexpr auto KMemoryRegionType_KernelMiscExceptionStack =
+    KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 3);
+constexpr auto KMemoryRegionType_KernelMiscUnknownDebug =
+    KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 4);
+// UNUSED: .Derive(7, 5);
+constexpr auto KMemoryRegionType_KernelMiscIdleStack =
+    KMemoryRegionType_KernelMiscDerivedBase.Derive(7, 6);
+static_assert(KMemoryRegionType_KernelMiscMainStack.GetValue() == 0xB49);
+static_assert(KMemoryRegionType_KernelMiscMappedDevice.GetValue() == 0xD49);
+static_assert(KMemoryRegionType_KernelMiscExceptionStack.GetValue() == 0x1349);
+static_assert(KMemoryRegionType_KernelMiscUnknownDebug.GetValue() == 0x1549);
+static_assert(KMemoryRegionType_KernelMiscIdleStack.GetValue() == 0x2349);
+
+constexpr auto KMemoryRegionType_KernelTemp = KMemoryRegionType_Kernel.Advance(2).Derive(2, 0);
+static_assert(KMemoryRegionType_KernelTemp.GetValue() == 0x31);
+
+constexpr KMemoryRegionType GetTypeForVirtualLinearMapping(u32 type_id) {
+    if (KMemoryRegionType_KernelTraceBuffer.IsAncestorOf(type_id)) {
+        return KMemoryRegionType_VirtualDramKernelTraceBuffer;
+    } else if (KMemoryRegionType_DramKernelPtHeap.IsAncestorOf(type_id)) {
+        return KMemoryRegionType_VirtualDramKernelPtHeap;
+    } else {
+        return KMemoryRegionType_Dram;
+    }
+}
+
+} // namespace Kernel

src/core/hle/kernel/k_page_bitmap.h

@@ -0,0 +1,279 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <bit>
+
+#include "common/alignment.h"
+#include "common/assert.h"
+#include "common/bit_util.h"
+#include "common/common_types.h"
+#include "common/tiny_mt.h"
+#include "core/hle/kernel/k_system_control.h"
+
+namespace Kernel {
+
+class KPageBitmap {
+private:
+    class RandomBitGenerator {
+    private:
+        Common::TinyMT rng{};
+        u32 entropy{};
+        u32 bits_available{};
+
+    private:
+        void RefreshEntropy() {
+            entropy = rng.GenerateRandomU32();
+            bits_available = static_cast<u32>(Common::BitSize<decltype(entropy)>());
+        }
+
+        bool GenerateRandomBit() {
+            if (bits_available == 0) {
+                this->RefreshEntropy();
+            }
+
+            const bool rnd_bit = (entropy & 1) != 0;
+            entropy >>= 1;
+            --bits_available;
+            return rnd_bit;
+        }
+
+    public:
+        RandomBitGenerator() {
+            rng.Initialize(static_cast<u32>(KSystemControl::GenerateRandomU64()));
+        }
+
+        std::size_t SelectRandomBit(u64 bitmap) {
+            u64 selected = 0;
+
+            u64 cur_num_bits = Common::BitSize<decltype(bitmap)>() / 2;
+            u64 cur_mask = (1ULL << cur_num_bits) - 1;
+
+            while (cur_num_bits) {
+                const u64 low = (bitmap >> 0) & cur_mask;
+                const u64 high = (bitmap >> cur_num_bits) & cur_mask;
+
+                bool choose_low;
+                if (high == 0) {
+                    // If only low val is set, choose low.
+                    choose_low = true;
+                } else if (low == 0) {
+                    // If only high val is set, choose high.
+                    choose_low = false;
+                } else {
+                    // If both are set, choose random.
+                    choose_low = this->GenerateRandomBit();
+                }
+
+                // If we chose low, proceed with low.
+                if (choose_low) {
+                    bitmap = low;
+                    selected += 0;
+                } else {
+                    bitmap = high;
+                    selected += cur_num_bits;
+                }
+
+                // Proceed.
+                cur_num_bits /= 2;
+                cur_mask >>= cur_num_bits;
+            }
+
+            return selected;
+        }
+    };
+
+public:
+    static constexpr std::size_t MaxDepth = 4;
+
+private:
+    std::array<u64*, MaxDepth> bit_storages{};
+    RandomBitGenerator rng{};
+    std::size_t num_bits{};
+    std::size_t used_depths{};
+
+public:
+    KPageBitmap() = default;
+
+    constexpr std::size_t GetNumBits() const {
+        return num_bits;
+    }
+    constexpr s32 GetHighestDepthIndex() const {
+        return static_cast<s32>(used_depths) - 1;
+    }
+
+    u64* Initialize(u64* storage, std::size_t size) {
+        // Initially, everything is un-set.
+        num_bits = 0;
+
+        // Calculate the needed bitmap depth.
+        used_depths = static_cast<std::size_t>(GetRequiredDepth(size));
+        ASSERT(used_depths <= MaxDepth);
+
+        // Set the bitmap pointers.
+        for (s32 depth = this->GetHighestDepthIndex(); depth >= 0; depth--) {
+            bit_storages[depth] = storage;
+            size = Common::AlignUp(size, Common::BitSize<u64>()) / Common::BitSize<u64>();
+            storage += size;
+        }
+
+        return storage;
+    }
+
+    s64 FindFreeBlock(bool random) {
+        uintptr_t offset = 0;
+        s32 depth = 0;
+
+        if (random) {
+            do {
+                const u64 v = bit_storages[depth][offset];
+                if (v == 0) {
+                    // If depth is bigger than zero, then a previous level indicated a block was
+                    // free.
+                    ASSERT(depth == 0);
+                    return -1;
+                }
+                offset = offset * Common::BitSize<u64>() + rng.SelectRandomBit(v);
+                ++depth;
+            } while (depth < static_cast<s32>(used_depths));
+        } else {
+            do {
+                const u64 v = bit_storages[depth][offset];
+                if (v == 0) {
+                    // If depth is bigger than zero, then a previous level indicated a block was
+                    // free.
+                    ASSERT(depth == 0);
+                    return -1;
+                }
+                offset = offset * Common::BitSize<u64>() + std::countr_zero(v);
+                ++depth;
+            } while (depth < static_cast<s32>(used_depths));
+        }
+
+        return static_cast<s64>(offset);
+    }
+
+    void SetBit(std::size_t offset) {
+        this->SetBit(this->GetHighestDepthIndex(), offset);
+        num_bits++;
+    }
+
+    void ClearBit(std::size_t offset) {
+        this->ClearBit(this->GetHighestDepthIndex(), offset);
+        num_bits--;
+    }
+
+    bool ClearRange(std::size_t offset, std::size_t count) {
+        s32 depth = this->GetHighestDepthIndex();
+        u64* bits = bit_storages[depth];
+        std::size_t bit_ind = offset / Common::BitSize<u64>();
+        if (count < Common::BitSize<u64>()) {
+            const std::size_t shift = offset % Common::BitSize<u64>();
+            ASSERT(shift + count <= Common::BitSize<u64>());
+            // Check that all the bits are set.
+            const u64 mask = ((u64(1) << count) - 1) << shift;
+            u64 v = bits[bit_ind];
+            if ((v & mask) != mask) {
+                return false;
+            }
+
+            // Clear the bits.
+            v &= ~mask;
+            bits[bit_ind] = v;
+            if (v == 0) {
+                this->ClearBit(depth - 1, bit_ind);
+            }
+        } else {
+            ASSERT(offset % Common::BitSize<u64>() == 0);
+            ASSERT(count % Common::BitSize<u64>() == 0);
+            // Check that all the bits are set.
+            std::size_t remaining = count;
+            std::size_t i = 0;
+            do {
+                if (bits[bit_ind + i++] != ~u64(0)) {
+                    return false;
+                }
+                remaining -= Common::BitSize<u64>();
+            } while (remaining > 0);
+
+            // Clear the bits.
+            remaining = count;
+            i = 0;
+            do {
+                bits[bit_ind + i] = 0;
+                this->ClearBit(depth - 1, bit_ind + i);
+                i++;
+                remaining -= Common::BitSize<u64>();
+            } while (remaining > 0);
+        }
+
+        num_bits -= count;
+        return true;
+    }
+
+private:
+    void SetBit(s32 depth, std::size_t offset) {
+        while (depth >= 0) {
+            std::size_t ind = offset / Common::BitSize<u64>();
+            std::size_t which = offset % Common::BitSize<u64>();
+            const u64 mask = u64(1) << which;
+
+            u64* bit = std::addressof(bit_storages[depth][ind]);
+            u64 v = *bit;
+            ASSERT((v & mask) == 0);
+            *bit = v | mask;
+            if (v) {
+                break;
+            }
+            offset = ind;
+            depth--;
+        }
+    }
+
+    void ClearBit(s32 depth, std::size_t offset) {
+        while (depth >= 0) {
+            std::size_t ind = offset / Common::BitSize<u64>();
+            std::size_t which = offset % Common::BitSize<u64>();
+            const u64 mask = u64(1) << which;
+
+            u64* bit = std::addressof(bit_storages[depth][ind]);
+            u64 v = *bit;
+            ASSERT((v & mask) != 0);
+            v &= ~mask;
+            *bit = v;
+            if (v) {
+                break;
+            }
+            offset = ind;
+            depth--;
+        }
+    }
+
+private:
+    static constexpr s32 GetRequiredDepth(std::size_t region_size) {
+        s32 depth = 0;
+        while (true) {
+            region_size /= Common::BitSize<u64>();
+            depth++;
+            if (region_size == 0) {
+                return depth;
+            }
+        }
+    }
+
+public:
+    static constexpr std::size_t CalculateManagementOverheadSize(std::size_t region_size) {
+        std::size_t overhead_bits = 0;
+        for (s32 depth = GetRequiredDepth(region_size) - 1; depth >= 0; depth--) {
+            region_size =
+                Common::AlignUp(region_size, Common::BitSize<u64>()) / Common::BitSize<u64>();
+            overhead_bits += region_size;
+        }
+        return overhead_bits * sizeof(u64);
+    }
+};
+
+} // namespace Kernel

src/core/hle/kernel/k_page_heap.cpp

@@ -2,16 +2,13 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-// This file references various implementation details from Atmosphere, an open-source firmware for
-// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX.
-
 #include "core/core.h"
-#include "core/hle/kernel/memory/page_heap.h"
+#include "core/hle/kernel/k_page_heap.h"
 #include "core/memory.h"
 
-namespace Kernel::Memory {
+namespace Kernel {
 
-void PageHeap::Initialize(VAddr address, std::size_t size, std::size_t metadata_size) {
+void KPageHeap::Initialize(VAddr address, std::size_t size, std::size_t metadata_size) {
     // Check our assumptions
     ASSERT(Common::IsAligned((address), PageSize));
     ASSERT(Common::IsAligned(size, PageSize));
@@ -32,11 +29,11 @@ void PageHeap::Initialize(VAddr address, std::size_t size, std::size_t metadata_
     }
 }
 
-VAddr PageHeap::AllocateBlock(s32 index) {
+VAddr KPageHeap::AllocateBlock(s32 index, bool random) {
     const std::size_t needed_size{blocks[index].GetSize()};
 
     for (s32 i{index}; i < static_cast<s32>(MemoryBlockPageShifts.size()); i++) {
-        if (const VAddr addr{blocks[i].PopBlock()}; addr) {
+        if (const VAddr addr{blocks[i].PopBlock(random)}; addr) {
             if (const std::size_t allocated_size{blocks[i].GetSize()};
                 allocated_size > needed_size) {
                 Free(addr + needed_size, (allocated_size - needed_size) / PageSize);
@@ -48,13 +45,13 @@ VAddr PageHeap::AllocateBlock(s32 index) {
     return 0;
 }
 
-void PageHeap::FreeBlock(VAddr block, s32 index) {
+void KPageHeap::FreeBlock(VAddr block, s32 index) {
     do {
         block = blocks[index++].PushBlock(block);
     } while (block != 0);
 }
 
-void PageHeap::Free(VAddr addr, std::size_t num_pages) {
+void KPageHeap::Free(VAddr addr, std::size_t num_pages) {
     // Freeing no pages is a no-op
     if (num_pages == 0) {
         return;
@@ -104,16 +101,16 @@ void PageHeap::Free(VAddr addr, std::size_t num_pages) {
     }
 }
 
-std::size_t PageHeap::CalculateMetadataOverheadSize(std::size_t region_size) {
+std::size_t KPageHeap::CalculateManagementOverheadSize(std::size_t region_size) {
     std::size_t overhead_size = 0;
     for (std::size_t i = 0; i < MemoryBlockPageShifts.size(); i++) {
         const std::size_t cur_block_shift{MemoryBlockPageShifts[i]};
         const std::size_t next_block_shift{
             (i != MemoryBlockPageShifts.size() - 1) ? MemoryBlockPageShifts[i + 1] : 0};
-        overhead_size += PageHeap::Block::CalculateMetadataOverheadSize(
+        overhead_size += KPageHeap::Block::CalculateManagementOverheadSize(
             region_size, cur_block_shift, next_block_shift);
     }
     return Common::AlignUp(overhead_size, PageSize);
 }
 
-} // namespace Kernel::Memory
+} // namespace Kernel

src/core/hle/kernel/k_page_heap.h

@@ -0,0 +1,193 @@
+// Copyright 2020 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <bit>
+#include <vector>
+
+#include "common/alignment.h"
+#include "common/assert.h"
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+#include "core/hle/kernel/k_page_bitmap.h"
+#include "core/hle/kernel/memory_types.h"
+
+namespace Kernel {
+
+class KPageHeap final : NonCopyable {
+public:
+    static constexpr s32 GetAlignedBlockIndex(std::size_t num_pages, std::size_t align_pages) {
+        const auto target_pages{std::max(num_pages, align_pages)};
+        for (std::size_t i = 0; i < NumMemoryBlockPageShifts; i++) {
+            if (target_pages <=
+                (static_cast<std::size_t>(1) << MemoryBlockPageShifts[i]) / PageSize) {
+                return static_cast<s32>(i);
+            }
+        }
+        return -1;
+    }
+
+    static constexpr s32 GetBlockIndex(std::size_t num_pages) {
+        for (s32 i{static_cast<s32>(NumMemoryBlockPageShifts) - 1}; i >= 0; i--) {
+            if (num_pages >= (static_cast<std::size_t>(1) << MemoryBlockPageShifts[i]) / PageSize) {
+                return i;
+            }
+        }
+        return -1;
+    }
+
+    static constexpr std::size_t GetBlockSize(std::size_t index) {
+        return static_cast<std::size_t>(1) << MemoryBlockPageShifts[index];
+    }
+
+    static constexpr std::size_t GetBlockNumPages(std::size_t index) {
+        return GetBlockSize(index) / PageSize;
+    }
+
+private:
+    static constexpr std::size_t NumMemoryBlockPageShifts{7};
+    static constexpr std::array<std::size_t, NumMemoryBlockPageShifts> MemoryBlockPageShifts{
+        0xC, 0x10, 0x15, 0x16, 0x19, 0x1D, 0x1E,
+    };
+
+    class Block final : NonCopyable {
+    private:
+        KPageBitmap bitmap;
+        VAddr heap_address{};
+        uintptr_t end_offset{};
+        std::size_t block_shift{};
+        std::size_t next_block_shift{};
+
+    public:
+        Block() = default;
+
+        constexpr std::size_t GetShift() const {
+            return block_shift;
+        }
+        constexpr std::size_t GetNextShift() const {
+            return next_block_shift;
+        }
+        constexpr std::size_t GetSize() const {
+            return static_cast<std::size_t>(1) << GetShift();
+        }
+        constexpr std::size_t GetNumPages() const {
+            return GetSize() / PageSize;
+        }
+        constexpr std::size_t GetNumFreeBlocks() const {
+            return bitmap.GetNumBits();
+        }
+        constexpr std::size_t GetNumFreePages() const {
+            return GetNumFreeBlocks() * GetNumPages();
+        }
+
+        u64* Initialize(VAddr addr, std::size_t size, std::size_t bs, std::size_t nbs,
+                        u64* bit_storage) {
+            // Set shifts
+            block_shift = bs;
+            next_block_shift = nbs;
+
+            // Align up the address
+            VAddr end{addr + size};
+            const auto align{(next_block_shift != 0) ? (1ULL << next_block_shift)
+                                                     : (1ULL << block_shift)};
+            addr = Common::AlignDown((addr), align);
+            end = Common::AlignUp((end), align);
+
+            heap_address = addr;
+            end_offset = (end - addr) / (1ULL << block_shift);
+            return bitmap.Initialize(bit_storage, end_offset);
+        }
+
+        VAddr PushBlock(VAddr address) {
+            // Set the bit for the free block
+            std::size_t offset{(address - heap_address) >> GetShift()};
+            bitmap.SetBit(offset);
+
+            // If we have a next shift, try to clear the blocks below and return the address
+            if (GetNextShift()) {
+                const auto diff{1ULL << (GetNextShift() - GetShift())};
+                offset = Common::AlignDown(offset, diff);
+                if (bitmap.ClearRange(offset, diff)) {
+                    return heap_address + (offset << GetShift());
+                }
+            }
+
+            // We couldn't coalesce, or we're already as big as possible
+            return 0;
+        }
+
+        VAddr PopBlock(bool random) {
+            // Find a free block
+            const s64 soffset{bitmap.FindFreeBlock(random)};
+            if (soffset < 0) {
+                return 0;
+            }
+            const auto offset{static_cast<std::size_t>(soffset)};
+
+            // Update our tracking and return it
+            bitmap.ClearBit(offset);
+            return heap_address + (offset << GetShift());
+        }
+
+    public:
+        static constexpr std::size_t CalculateManagementOverheadSize(std::size_t region_size,
+                                                                     std::size_t cur_block_shift,
+                                                                     std::size_t next_block_shift) {
+            const auto cur_block_size{(1ULL << cur_block_shift)};
+            const auto next_block_size{(1ULL << next_block_shift)};
+            const auto align{(next_block_shift != 0) ? next_block_size : cur_block_size};
+            return KPageBitmap::CalculateManagementOverheadSize(
+                (align * 2 + Common::AlignUp(region_size, align)) / cur_block_size);
+        }
+    };
+
+public:
+    KPageHeap() = default;
+
+    constexpr VAddr GetAddress() const {
+        return heap_address;
+    }
+    constexpr std::size_t GetSize() const {
+        return heap_size;
+    }
+    constexpr VAddr GetEndAddress() const {
+        return GetAddress() + GetSize();
+    }
+    constexpr std::size_t GetPageOffset(VAddr block) const {
+        return (block - GetAddress()) / PageSize;
+    }
+
+    void Initialize(VAddr heap_address, std::size_t heap_size, std::size_t metadata_size);
+    VAddr AllocateBlock(s32 index, bool random);
+    void Free(VAddr addr, std::size_t num_pages);
+
+    void UpdateUsedSize() {
+        used_size = heap_size - (GetNumFreePages() * PageSize);
+    }
+
+    static std::size_t CalculateManagementOverheadSize(std::size_t region_size);
+
+private:
+    constexpr std::size_t GetNumFreePages() const {
+        std::size_t num_free{};
+
+        for (const auto& block : blocks) {
+            num_free += block.GetNumFreePages();
+        }
+
+        return num_free;
+    }
+
+    void FreeBlock(VAddr block, s32 index);
+
+    VAddr heap_address{};
+    std::size_t heap_size{};
+    std::size_t used_size{};
+    std::array<Block, NumMemoryBlockPageShifts> blocks{};
+    std::vector<u64> metadata;
+};
+
+} // namespace Kernel

src/core/hle/kernel/k_page_linked_list.h

@@ -8,12 +8,12 @@
 
 #include "common/assert.h"
 #include "common/common_types.h"
-#include "core/hle/kernel/memory/memory_types.h"
+#include "core/hle/kernel/memory_types.h"
 #include "core/hle/result.h"
 
-namespace Kernel::Memory {
+namespace Kernel {
 
-class PageLinkedList final {
+class KPageLinkedList final {
 public:
     class Node final {
     public:
@@ -33,8 +33,8 @@ public:
     };
 
 public:
-    PageLinkedList() = default;
-    PageLinkedList(u64 address, u64 num_pages) {
+    KPageLinkedList() = default;
+    KPageLinkedList(u64 address, u64 num_pages) {
         ASSERT(AddBlock(address, num_pages).IsSuccess());
     }
 
@@ -54,7 +54,7 @@ public:
         return num_pages;
     }
 
-    bool IsEqual(PageLinkedList& other) const {
+    bool IsEqual(KPageLinkedList& other) const {
         auto this_node = nodes.begin();
         auto other_node = other.nodes.begin();
         while (this_node != nodes.end() && other_node != other.nodes.end()) {
@@ -89,4 +89,4 @@ private:
     std::list<Node> nodes;
 };
 
-} // namespace Kernel::Memory
+} // namespace Kernel

src/core/hle/kernel/k_page_table.h

@@ -10,27 +10,27 @@
 #include "common/common_types.h"
 #include "common/page_table.h"
 #include "core/file_sys/program_metadata.h"
-#include "core/hle/kernel/memory/memory_block.h"
-#include "core/hle/kernel/memory/memory_manager.h"
+#include "core/hle/kernel/k_memory_block.h"
+#include "core/hle/kernel/k_memory_manager.h"
 #include "core/hle/result.h"
 
 namespace Core {
 class System;
 }
 
-namespace Kernel::Memory {
+namespace Kernel {
 
-class MemoryBlockManager;
+class KMemoryBlockManager;
 
-class PageTable final : NonCopyable {
+class KPageTable final : NonCopyable {
 public:
-    explicit PageTable(Core::System& system);
+    explicit KPageTable(Core::System& system);
 
     ResultCode InitializeForProcess(FileSys::ProgramAddressSpaceType as_type, bool enable_aslr,
                                     VAddr code_addr, std::size_t code_size,
-                                    Memory::MemoryManager::Pool pool);
-    ResultCode MapProcessCode(VAddr addr, std::size_t pages_count, MemoryState state,
-                              MemoryPermission perm);
+                                    KMemoryManager::Pool pool);
+    ResultCode MapProcessCode(VAddr addr, std::size_t pages_count, KMemoryState state,
+                              KMemoryPermission perm);
     ResultCode MapProcessCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
     ResultCode UnmapProcessCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size);
     ResultCode MapPhysicalMemory(VAddr addr, std::size_t size);
@@ -38,20 +38,20 @@ public:
     ResultCode UnmapMemory(VAddr addr, std::size_t size);
     ResultCode Map(VAddr dst_addr, VAddr src_addr, std::size_t size);
     ResultCode Unmap(VAddr dst_addr, VAddr src_addr, std::size_t size);
-    ResultCode MapPages(VAddr addr, PageLinkedList& page_linked_list, MemoryState state,
-                        MemoryPermission perm);
-    ResultCode SetCodeMemoryPermission(VAddr addr, std::size_t size, MemoryPermission perm);
-    MemoryInfo QueryInfo(VAddr addr);
-    ResultCode ReserveTransferMemory(VAddr addr, std::size_t size, MemoryPermission perm);
+    ResultCode MapPages(VAddr addr, KPageLinkedList& page_linked_list, KMemoryState state,
+                        KMemoryPermission perm);
+    ResultCode SetCodeMemoryPermission(VAddr addr, std::size_t size, KMemoryPermission perm);
+    KMemoryInfo QueryInfo(VAddr addr);
+    ResultCode ReserveTransferMemory(VAddr addr, std::size_t size, KMemoryPermission perm);
     ResultCode ResetTransferMemory(VAddr addr, std::size_t size);
-    ResultCode SetMemoryAttribute(VAddr addr, std::size_t size, MemoryAttribute mask,
-                                  MemoryAttribute value);
+    ResultCode SetMemoryAttribute(VAddr addr, std::size_t size, KMemoryAttribute mask,
+                                  KMemoryAttribute value);
     ResultCode SetHeapCapacity(std::size_t new_heap_capacity);
     ResultVal<VAddr> SetHeapSize(std::size_t size);
     ResultVal<VAddr> AllocateAndMapMemory(std::size_t needed_num_pages, std::size_t align,
                                           bool is_map_only, VAddr region_start,
-                                          std::size_t region_num_pages, MemoryState state,
-                                          MemoryPermission perm, PAddr map_addr = 0);
+                                          std::size_t region_num_pages, KMemoryState state,
+                                          KMemoryPermission perm, PAddr map_addr = 0);
     ResultCode LockForDeviceAddressSpace(VAddr addr, std::size_t size);
     ResultCode UnlockForDeviceAddressSpace(VAddr addr, std::size_t size);
 
@@ -72,47 +72,49 @@ private:
         ChangePermissionsAndRefresh,
     };
 
-    static constexpr MemoryAttribute DefaultMemoryIgnoreAttr =
-        MemoryAttribute::DontCareMask | MemoryAttribute::IpcLocked | MemoryAttribute::DeviceShared;
+    static constexpr KMemoryAttribute DefaultMemoryIgnoreAttr = KMemoryAttribute::DontCareMask |
+                                                                KMemoryAttribute::IpcLocked |
+                                                                KMemoryAttribute::DeviceShared;
 
     ResultCode InitializeMemoryLayout(VAddr start, VAddr end);
-    ResultCode MapPages(VAddr addr, const PageLinkedList& page_linked_list, MemoryPermission perm);
-    void MapPhysicalMemory(PageLinkedList& page_linked_list, VAddr start, VAddr end);
+    ResultCode MapPages(VAddr addr, const KPageLinkedList& page_linked_list,
+                        KMemoryPermission perm);
+    void MapPhysicalMemory(KPageLinkedList& page_linked_list, VAddr start, VAddr end);
     bool IsRegionMapped(VAddr address, u64 size);
     bool IsRegionContiguous(VAddr addr, u64 size) const;
-    void AddRegionToPages(VAddr start, std::size_t num_pages, PageLinkedList& page_linked_list);
-    MemoryInfo QueryInfoImpl(VAddr addr);
+    void AddRegionToPages(VAddr start, std::size_t num_pages, KPageLinkedList& page_linked_list);
+    KMemoryInfo QueryInfoImpl(VAddr addr);
     VAddr AllocateVirtualMemory(VAddr start, std::size_t region_num_pages, u64 needed_num_pages,
                                 std::size_t align);
-    ResultCode Operate(VAddr addr, std::size_t num_pages, const PageLinkedList& page_group,
+    ResultCode Operate(VAddr addr, std::size_t num_pages, const KPageLinkedList& page_group,
                        OperationType operation);
-    ResultCode Operate(VAddr addr, std::size_t num_pages, MemoryPermission perm,
+    ResultCode Operate(VAddr addr, std::size_t num_pages, KMemoryPermission perm,
                        OperationType operation, PAddr map_addr = 0);
-    constexpr VAddr GetRegionAddress(MemoryState state) const;
-    constexpr std::size_t GetRegionSize(MemoryState state) const;
-    constexpr bool CanContain(VAddr addr, std::size_t size, MemoryState state) const;
+    constexpr VAddr GetRegionAddress(KMemoryState state) const;
+    constexpr std::size_t GetRegionSize(KMemoryState state) const;
+    constexpr bool CanContain(VAddr addr, std::size_t size, KMemoryState state) const;
 
-    constexpr ResultCode CheckMemoryState(const MemoryInfo& info, MemoryState state_mask,
-                                          MemoryState state, MemoryPermission perm_mask,
-                                          MemoryPermission perm, MemoryAttribute attr_mask,
-                                          MemoryAttribute attr) const;
-    ResultCode CheckMemoryState(MemoryState* out_state, MemoryPermission* out_perm,
-                                MemoryAttribute* out_attr, VAddr addr, std::size_t size,
-                                MemoryState state_mask, MemoryState state,
-                                MemoryPermission perm_mask, MemoryPermission perm,
-                                MemoryAttribute attr_mask, MemoryAttribute attr,
-                                MemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr);
-    ResultCode CheckMemoryState(VAddr addr, std::size_t size, MemoryState state_mask,
-                                MemoryState state, MemoryPermission perm_mask,
-                                MemoryPermission perm, MemoryAttribute attr_mask,
-                                MemoryAttribute attr,
-                                MemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr) {
+    constexpr ResultCode CheckMemoryState(const KMemoryInfo& info, KMemoryState state_mask,
+                                          KMemoryState state, KMemoryPermission perm_mask,
+                                          KMemoryPermission perm, KMemoryAttribute attr_mask,
+                                          KMemoryAttribute attr) const;
+    ResultCode CheckMemoryState(KMemoryState* out_state, KMemoryPermission* out_perm,
+                                KMemoryAttribute* out_attr, VAddr addr, std::size_t size,
+                                KMemoryState state_mask, KMemoryState state,
+                                KMemoryPermission perm_mask, KMemoryPermission perm,
+                                KMemoryAttribute attr_mask, KMemoryAttribute attr,
+                                KMemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr);
+    ResultCode CheckMemoryState(VAddr addr, std::size_t size, KMemoryState state_mask,
+                                KMemoryState state, KMemoryPermission perm_mask,
+                                KMemoryPermission perm, KMemoryAttribute attr_mask,
+                                KMemoryAttribute attr,
+                                KMemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr) {
         return CheckMemoryState(nullptr, nullptr, nullptr, addr, size, state_mask, state, perm_mask,
                                 perm, attr_mask, attr, ignore_attr);
     }
 
     std::recursive_mutex page_table_lock;
-    std::unique_ptr<MemoryBlockManager> block_manager;
+    std::unique_ptr<KMemoryBlockManager> block_manager;
 
 public:
     constexpr VAddr GetAddressSpaceStart() const {
@@ -212,7 +214,7 @@ public:
         return !IsOutsideASLRRegion(address, size);
     }
     constexpr PAddr GetPhysicalAddr(VAddr addr) {
-        return page_table_impl.backing_addr[addr >> Memory::PageBits] + addr;
+        return page_table_impl.backing_addr[addr >> PageBits] + addr;
     }
 
 private:
@@ -267,11 +269,11 @@ private:
     bool is_kernel{};
     bool is_aslr_enabled{};
 
-    MemoryManager::Pool memory_pool{MemoryManager::Pool::Application};
+    KMemoryManager::Pool memory_pool{KMemoryManager::Pool::Application};
 
     Common::PageTable page_table_impl;
 
     Core::System& system;
 };
 
-} // namespace Kernel::Memory
+} // namespace Kernel

src/core/hle/kernel/k_scheduler.cpp

@@ -62,7 +62,7 @@ void KScheduler::RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedul
 }
 
 u64 KScheduler::UpdateHighestPriorityThread(KThread* highest_thread) {
-    std::scoped_lock lock{guard};
+    KScopedSpinLock lk{guard};
     if (KThread* prev_highest_thread = state.highest_priority_thread;
         prev_highest_thread != highest_thread) {
         if (prev_highest_thread != nullptr) {
@@ -637,11 +637,11 @@ void KScheduler::RescheduleCurrentCore() {
     if (phys_core.IsInterrupted()) {
         phys_core.ClearInterrupt();
     }
-    guard.lock();
+    guard.Lock();
     if (state.needs_scheduling.load()) {
         Schedule();
     } else {
-        guard.unlock();
+        guard.Unlock();
     }
 }
 
@@ -669,7 +669,7 @@ void KScheduler::Unload(KThread* thread) {
         } else {
             prev_thread = nullptr;
         }
-        thread->context_guard.unlock();
+        thread->context_guard.Unlock();
     }
 }
 
@@ -713,7 +713,7 @@ void KScheduler::ScheduleImpl() {
 
     // If we're not actually switching thread, there's nothing to do.
     if (next_thread == current_thread.load()) {
-        guard.unlock();
+        guard.Unlock();
         return;
     }
 
@@ -732,9 +732,9 @@ void KScheduler::ScheduleImpl() {
     } else {
         old_context = &idle_thread->GetHostContext();
     }
-    guard.unlock();
+    guard.Unlock();
 
-    Common::Fiber::YieldTo(*old_context, switch_fiber);
+    Common::Fiber::YieldTo(*old_context, *switch_fiber);
     /// When a thread wakes up, the scheduler may have changed to other in another core.
     auto& next_scheduler = *system.Kernel().CurrentScheduler();
     next_scheduler.SwitchContextStep2();
@@ -748,34 +748,29 @@ void KScheduler::OnSwitch(void* this_scheduler) {
 void KScheduler::SwitchToCurrent() {
     while (true) {
         {
-            std::scoped_lock lock{guard};
+            KScopedSpinLock lk{guard};
             current_thread.store(state.highest_priority_thread);
             state.needs_scheduling.store(false);
         }
         const auto is_switch_pending = [this] {
-            std::scoped_lock lock{guard};
+            KScopedSpinLock lk{guard};
             return state.needs_scheduling.load();
         };
         do {
             auto next_thread = current_thread.load();
             if (next_thread != nullptr) {
-                next_thread->context_guard.lock();
+                next_thread->context_guard.Lock();
                 if (next_thread->GetRawState() != ThreadState::Runnable) {
-                    next_thread->context_guard.unlock();
+                    next_thread->context_guard.Unlock();
                     break;
                 }
                 if (next_thread->GetActiveCore() != core_id) {
-                    next_thread->context_guard.unlock();
+                    next_thread->context_guard.Unlock();
                     break;
                 }
             }
-            std::shared_ptr<Common::Fiber>* next_context;
-            if (next_thread != nullptr) {
-                next_context = &next_thread->GetHostContext();
-            } else {
-                next_context = &idle_thread->GetHostContext();
-            }
-            Common::Fiber::YieldTo(switch_fiber, *next_context);
+            auto thread = next_thread ? next_thread : idle_thread;
+            Common::Fiber::YieldTo(switch_fiber, *thread->GetHostContext());
         } while (!is_switch_pending());
     }
 }
@@ -800,9 +795,9 @@ void KScheduler::Initialize() {
     std::string name = "Idle Thread Id:" + std::to_string(core_id);
     std::function<void(void*)> init_func = Core::CpuManager::GetIdleThreadStartFunc();
     void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater();
-    auto thread_res = KThread::Create(system, ThreadType::Main, name, 0,
-                                      KThread::IdleThreadPriority, 0, static_cast<u32>(core_id), 0,
-                                      nullptr, std::move(init_func), init_func_parameter);
+    auto thread_res = KThread::CreateThread(
+        system, ThreadType::Main, name, 0, KThread::IdleThreadPriority, 0,
+        static_cast<u32>(core_id), 0, nullptr, std::move(init_func), init_func_parameter);
     idle_thread = thread_res.Unwrap().get();
 }
 

src/core/hle/kernel/k_scheduler.h

@@ -2,19 +2,16 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-// This file references various implementation details from Atmosphere, an open-source firmware for
-// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX.
-
 #pragma once
 
 #include <atomic>
 
 #include "common/common_types.h"
-#include "common/spin_lock.h"
 #include "core/hle/kernel/global_scheduler_context.h"
 #include "core/hle/kernel/k_priority_queue.h"
 #include "core/hle/kernel/k_scheduler_lock.h"
 #include "core/hle/kernel/k_scoped_lock.h"
+#include "core/hle/kernel/k_spin_lock.h"
 
 namespace Common {
 class Fiber;
@@ -195,12 +192,12 @@ private:
     u64 last_context_switch_time{};
     const s32 core_id;
 
-    Common::SpinLock guard{};
+    KSpinLock guard{};
 };
 
-class KScopedSchedulerLock : KScopedLock<GlobalSchedulerContext::LockType> {
+class [[nodiscard]] KScopedSchedulerLock : KScopedLock<GlobalSchedulerContext::LockType> {
 public:
-    explicit KScopedSchedulerLock(KernelCore& kernel);
+    explicit KScopedSchedulerLock(KernelCore & kernel);
     ~KScopedSchedulerLock();
 };
 

src/core/hle/kernel/k_scheduler_lock.h

@@ -2,14 +2,11 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-// This file references various implementation details from Atmosphere, an open-source firmware for
-// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX.
-
 #pragma once
 
 #include "common/assert.h"
-#include "common/spin_lock.h"
 #include "core/hardware_properties.h"
+#include "core/hle/kernel/k_spin_lock.h"
 #include "core/hle/kernel/k_thread.h"
 #include "core/hle/kernel/kernel.h"
 
@@ -34,7 +31,7 @@ public:
         } else {
             // Otherwise, we want to disable scheduling and acquire the spinlock.
             SchedulerType::DisableScheduling(kernel);
-            spin_lock.lock();
+            spin_lock.Lock();
 
             // For debug, ensure that our state is valid.
             ASSERT(lock_count == 0);
@@ -58,7 +55,7 @@ public:
 
             // Note that we no longer hold the lock, and unlock the spinlock.
             owner_thread = nullptr;
-            spin_lock.unlock();
+            spin_lock.Unlock();
 
             // Enable scheduling, and perform a rescheduling operation.
             SchedulerType::EnableScheduling(kernel, cores_needing_scheduling);
@@ -67,7 +64,7 @@ public:
 
 private:
     KernelCore& kernel;
-    Common::SpinLock spin_lock{};
+    KAlignedSpinLock spin_lock{};
     s32 lock_count{};
     KThread* owner_thread{};
 };

src/core/hle/kernel/k_scoped_lock.h

@@ -20,19 +20,22 @@ concept KLockable = !std::is_reference_v<T> && requires(T & t) {
 };
 
 template <typename T>
-requires KLockable<T> class KScopedLock {
+requires KLockable<T> class [[nodiscard]] KScopedLock {
 public:
-    explicit KScopedLock(T* l) : lock_ptr(l) {
+    explicit KScopedLock(T * l) : lock_ptr(l) {
         this->lock_ptr->Lock();
     }
-    explicit KScopedLock(T& l) : KScopedLock(std::addressof(l)) { /* ... */
-    }
+    explicit KScopedLock(T & l) : KScopedLock(std::addressof(l)) {}
+
     ~KScopedLock() {
         this->lock_ptr->Unlock();
     }
 
     KScopedLock(const KScopedLock&) = delete;
-    KScopedLock(KScopedLock&&) = delete;
+    KScopedLock& operator=(const KScopedLock&) = delete;
+
+    KScopedLock(KScopedLock &&) = delete;
+    KScopedLock& operator=(KScopedLock&&) = delete;
 
 private:
     T* lock_ptr;

src/core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h

@@ -15,9 +15,9 @@
 
 namespace Kernel {
 
-class KScopedSchedulerLockAndSleep {
+class [[nodiscard]] KScopedSchedulerLockAndSleep {
 public:
-    explicit KScopedSchedulerLockAndSleep(KernelCore& kernel, KThread* t, s64 timeout)
+    explicit KScopedSchedulerLockAndSleep(KernelCore & kernel, KThread * t, s64 timeout)
         : kernel(kernel), thread(t), timeout_tick(timeout) {
         // Lock the scheduler.
         kernel.GlobalSchedulerContext().scheduler_lock.Lock();

src/core/hle/kernel/k_shared_memory.cpp

@@ -4,33 +4,32 @@
 
 #include "common/assert.h"
 #include "core/core.h"
+#include "core/hle/kernel/k_page_table.h"
 #include "core/hle/kernel/k_scoped_resource_reservation.h"
+#include "core/hle/kernel/k_shared_memory.h"
 #include "core/hle/kernel/kernel.h"
-#include "core/hle/kernel/memory/page_table.h"
-#include "core/hle/kernel/shared_memory.h"
 
 namespace Kernel {
 
-SharedMemory::SharedMemory(KernelCore& kernel, Core::DeviceMemory& device_memory)
+KSharedMemory::KSharedMemory(KernelCore& kernel, Core::DeviceMemory& device_memory)
     : Object{kernel}, device_memory{device_memory} {}
 
-SharedMemory::~SharedMemory() {
+KSharedMemory::~KSharedMemory() {
     kernel.GetSystemResourceLimit()->Release(LimitableResource::PhysicalMemory, size);
 }
 
-std::shared_ptr<SharedMemory> SharedMemory::Create(
+std::shared_ptr<KSharedMemory> KSharedMemory::Create(
     KernelCore& kernel, Core::DeviceMemory& device_memory, Process* owner_process,
-    Memory::PageLinkedList&& page_list, Memory::MemoryPermission owner_permission,
-    Memory::MemoryPermission user_permission, PAddr physical_address, std::size_t size,
-    std::string name) {
+    KPageLinkedList&& page_list, KMemoryPermission owner_permission,
+    KMemoryPermission user_permission, PAddr physical_address, std::size_t size, std::string name) {
 
     const auto resource_limit = kernel.GetSystemResourceLimit();
     KScopedResourceReservation memory_reservation(resource_limit, LimitableResource::PhysicalMemory,
                                                   size);
     ASSERT(memory_reservation.Succeeded());
 
-    std::shared_ptr<SharedMemory> shared_memory{
-        std::make_shared<SharedMemory>(kernel, device_memory)};
+    std::shared_ptr<KSharedMemory> shared_memory{
+        std::make_shared<KSharedMemory>(kernel, device_memory)};
 
     shared_memory->owner_process = owner_process;
     shared_memory->page_list = std::move(page_list);
@@ -44,22 +43,22 @@ std::shared_ptr<SharedMemory> SharedMemory::Create(
     return shared_memory;
 }
 
-ResultCode SharedMemory::Map(Process& target_process, VAddr address, std::size_t size,
-                             Memory::MemoryPermission permissions) {
-    const u64 page_count{(size + Memory::PageSize - 1) / Memory::PageSize};
+ResultCode KSharedMemory::Map(Process& target_process, VAddr address, std::size_t size,
+                              KMemoryPermission permissions) {
+    const u64 page_count{(size + PageSize - 1) / PageSize};
 
     if (page_list.GetNumPages() != page_count) {
         UNIMPLEMENTED_MSG("Page count does not match");
     }
 
-    const Memory::MemoryPermission expected =
+    const KMemoryPermission expected =
         &target_process == owner_process ? owner_permission : user_permission;
 
     if (permissions != expected) {
         UNIMPLEMENTED_MSG("Permission does not match");
     }
 
-    return target_process.PageTable().MapPages(address, page_list, Memory::MemoryState::Shared,
+    return target_process.PageTable().MapPages(address, page_list, KMemoryState::Shared,
                                                permissions);
 }
 

src/core/hle/kernel/k_shared_memory.h

@@ -9,8 +9,8 @@
 
 #include "common/common_types.h"
 #include "core/device_memory.h"
-#include "core/hle/kernel/memory/memory_block.h"
-#include "core/hle/kernel/memory/page_linked_list.h"
+#include "core/hle/kernel/k_memory_block.h"
+#include "core/hle/kernel/k_page_linked_list.h"
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/result.h"
@@ -19,15 +19,15 @@ namespace Kernel {
 
 class KernelCore;
 
-class SharedMemory final : public Object {
+class KSharedMemory final : public Object {
 public:
-    explicit SharedMemory(KernelCore& kernel, Core::DeviceMemory& device_memory);
-    ~SharedMemory() override;
+    explicit KSharedMemory(KernelCore& kernel, Core::DeviceMemory& device_memory);
+    ~KSharedMemory() override;
 
-    static std::shared_ptr<SharedMemory> Create(
+    static std::shared_ptr<KSharedMemory> Create(
         KernelCore& kernel, Core::DeviceMemory& device_memory, Process* owner_process,
-        Memory::PageLinkedList&& page_list, Memory::MemoryPermission owner_permission,
-        Memory::MemoryPermission user_permission, PAddr physical_address, std::size_t size,
+        KPageLinkedList&& page_list, KMemoryPermission owner_permission,
+        KMemoryPermission user_permission, PAddr physical_address, std::size_t size,
         std::string name);
 
     std::string GetTypeName() const override {
@@ -51,7 +51,7 @@ public:
      * @param permissions Memory block map permissions (specified by SVC field)
      */
     ResultCode Map(Process& target_process, VAddr address, std::size_t size,
-                   Memory::MemoryPermission permissions);
+                   KMemoryPermission permissions);
 
     /**
      * Gets a pointer to the shared memory block
@@ -76,9 +76,9 @@ public:
 private:
     Core::DeviceMemory& device_memory;
     Process* owner_process{};
-    Memory::PageLinkedList page_list;
-    Memory::MemoryPermission owner_permission{};
-    Memory::MemoryPermission user_permission{};
+    KPageLinkedList page_list;
+    KMemoryPermission owner_permission{};
+    KMemoryPermission user_permission{};
     PAddr physical_address{};
     std::size_t size{};
     std::string name;

src/core/hle/kernel/k_slab_heap.h

@@ -2,9 +2,6 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-// This file references various implementation details from Atmosphere, an open-source firmware for
-// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX.
-
 #pragma once
 
 #include <atomic>
@@ -12,17 +9,17 @@
 #include "common/assert.h"
 #include "common/common_types.h"
 
-namespace Kernel::Memory {
+namespace Kernel {
 
 namespace impl {
 
-class SlabHeapImpl final : NonCopyable {
+class KSlabHeapImpl final : NonCopyable {
 public:
     struct Node {
         Node* next{};
     };
 
-    constexpr SlabHeapImpl() = default;
+    constexpr KSlabHeapImpl() = default;
 
     void Initialize(std::size_t size) {
         ASSERT(head == nullptr);
@@ -65,9 +62,9 @@ private:
 
 } // namespace impl
 
-class SlabHeapBase : NonCopyable {
+class KSlabHeapBase : NonCopyable {
 public:
-    constexpr SlabHeapBase() = default;
+    constexpr KSlabHeapBase() = default;
 
     constexpr bool Contains(uintptr_t addr) const {
         return start <= addr && addr < end;
@@ -126,7 +123,7 @@ public:
     }
 
 private:
-    using Impl = impl::SlabHeapImpl;
+    using Impl = impl::KSlabHeapImpl;
 
     Impl impl;
     uintptr_t peak{};
@@ -135,9 +132,9 @@ private:
 };
 
 template <typename T>
-class SlabHeap final : public SlabHeapBase {
+class KSlabHeap final : public KSlabHeapBase {
 public:
-    constexpr SlabHeap() : SlabHeapBase() {}
+    constexpr KSlabHeap() : KSlabHeapBase() {}
 
     void Initialize(void* memory, std::size_t memory_size) {
         InitializeImpl(sizeof(T), memory, memory_size);
@@ -160,4 +157,4 @@ public:
     }
 };
 
-} // namespace Kernel::Memory
+} // namespace Kernel

src/core/hle/kernel/k_spin_lock.cpp

@@ -0,0 +1,54 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/hle/kernel/k_spin_lock.h"
+
+#if _MSC_VER
+#include <intrin.h>
+#if _M_AMD64
+#define __x86_64__ 1
+#endif
+#if _M_ARM64
+#define __aarch64__ 1
+#endif
+#else
+#if __x86_64__
+#include <xmmintrin.h>
+#endif
+#endif
+
+namespace {
+
+void ThreadPause() {
+#if __x86_64__
+    _mm_pause();
+#elif __aarch64__ && _MSC_VER
+    __yield();
+#elif __aarch64__
+    asm("yield");
+#endif
+}
+
+} // namespace
+
+namespace Kernel {
+
+void KSpinLock::Lock() {
+    while (lck.test_and_set(std::memory_order_acquire)) {
+        ThreadPause();
+    }
+}
+
+void KSpinLock::Unlock() {
+    lck.clear(std::memory_order_release);
+}
+
+bool KSpinLock::TryLock() {
+    if (lck.test_and_set(std::memory_order_acquire)) {
+        return false;
+    }
+    return true;
+}
+
+} // namespace Kernel

src/core/hle/kernel/k_spin_lock.h

@@ -0,0 +1,39 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <atomic>
+
+#include "core/hle/kernel/k_scoped_lock.h"
+
+namespace Kernel {
+
+class KSpinLock {
+public:
+    KSpinLock() = default;
+
+    KSpinLock(const KSpinLock&) = delete;
+    KSpinLock& operator=(const KSpinLock&) = delete;
+
+    KSpinLock(KSpinLock&&) = delete;
+    KSpinLock& operator=(KSpinLock&&) = delete;
+
+    void Lock();
+    void Unlock();
+    [[nodiscard]] bool TryLock();
+
+private:
+    std::atomic_flag lck = ATOMIC_FLAG_INIT;
+};
+
+// TODO(bunnei): Alias for now, in case we want to implement these accurately in the future.
+using KAlignedSpinLock = KSpinLock;
+using KNotAlignedSpinLock = KSpinLock;
+
+using KScopedSpinLock = KScopedLock<KSpinLock>;
+using KScopedAlignedSpinLock = KScopedLock<KAlignedSpinLock>;
+using KScopedNotAlignedSpinLock = KScopedLock<KNotAlignedSpinLock>;
+
+} // namespace Kernel

src/core/hle/kernel/k_system_control.h

@@ -0,0 +1,23 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+
+#define BOARD_NINTENDO_NX
+
+#ifdef BOARD_NINTENDO_NX
+
+#include "core/hle/kernel/board/nintendo/nx/k_system_control.h"
+
+namespace Kernel {
+
+using Kernel::Board::Nintendo::Nx::KSystemControl;
+
+} // namespace Kernel
+
+#else
+#error "Unknown board for KSystemControl"
+#endif