externals: Update httplib

Since the introduction of this library, numerous improvements have been
made. Notably, many of the warnings we would get by simply including the
library header have now been fixed. This makes it much easier to make
conversion warning an error.

.ci/scripts/common/post-upload.sh

@@ -1,12 +1,12 @@
 #!/bin/bash -ex
 
 # Copy documentation
-cp license.txt "$REV_NAME"
-cp README.md "$REV_NAME"
+cp license.txt "$DIR_NAME"
+cp README.md "$DIR_NAME"
 
-tar $COMPRESSION_FLAGS "$ARCHIVE_NAME" "$REV_NAME"
+tar $COMPRESSION_FLAGS "$ARCHIVE_NAME" "$DIR_NAME"
 
-mv "$REV_NAME" $RELEASE_NAME
+mv "$DIR_NAME" $RELEASE_NAME
 
 7z a "$REV_NAME.7z" $RELEASE_NAME
 

.ci/scripts/linux/docker.sh

@@ -11,5 +11,4 @@ ninja
 
 ccache -s
 
-# Ignore zlib's tests, since they aren't gated behind a CMake option.
-ctest -VV -E "(example|example64)" -C Release
+ctest -VV -C Release

.ci/scripts/linux/upload.sh

@@ -6,9 +6,15 @@ REV_NAME="yuzu-linux-${GITDATE}-${GITREV}"
 ARCHIVE_NAME="${REV_NAME}.tar.xz"
 COMPRESSION_FLAGS="-cJvf"
 
-mkdir "$REV_NAME"
+if [ "${RELEASE_NAME}" = "mainline" ]; then
+    DIR_NAME="${REV_NAME}"
+else
+    DIR_NAME="${REV_NAME}_${RELEASE_NAME}"
+fi
 
-cp build/bin/yuzu-cmd "$REV_NAME"
-cp build/bin/yuzu "$REV_NAME"
+mkdir "$DIR_NAME"
+
+cp build/bin/yuzu-cmd "$DIR_NAME"
+cp build/bin/yuzu "$DIR_NAME"
 
 . .ci/scripts/common/post-upload.sh

.ci/scripts/windows/upload.ps1

@@ -1,6 +1,13 @@
+param($BUILD_NAME)
+
 $GITDATE = $(git show -s --date=short --format='%ad') -replace "-",""
 $GITREV = $(git show -s --format='%h')
-$RELEASE_DIST = "yuzu-windows-msvc"
+
+if ("$BUILD_NAME" -eq "mainline") {
+    $RELEASE_DIST = "yuzu-windows-msvc"
+} else {
+    $RELEASE_DIST = "yuzu-windows-msvc-$BUILD_NAME"
+}
 
 $MSVC_BUILD_ZIP = "yuzu-windows-msvc-$GITDATE-$GITREV.zip" -replace " ", ""
 $MSVC_BUILD_PDB = "yuzu-windows-msvc-$GITDATE-$GITREV-debugsymbols.zip" -replace " ", ""

.ci/scripts/windows/upload.sh

@@ -6,8 +6,14 @@ REV_NAME="yuzu-windows-mingw-${GITDATE}-${GITREV}"
 ARCHIVE_NAME="${REV_NAME}.tar.gz"
 COMPRESSION_FLAGS="-czvf"
 
-mkdir "$REV_NAME"
+if [ "${RELEASE_NAME}" = "mainline" ]; then
+    DIR_NAME="${REV_NAME}"
+else
+    DIR_NAME="${REV_NAME}_${RELEASE_NAME}"
+fi
+
+mkdir "$DIR_NAME"
 # get around the permission issues
-cp -r package/* "$REV_NAME"
+cp -r package/* "$DIR_NAME"
 
 . .ci/scripts/common/post-upload.sh

.ci/templates/build-msvc.yml

@@ -17,6 +17,7 @@ steps:
   inputs:
     targetType: 'filePath'
     filePath: './.ci/scripts/windows/upload.ps1'
+    arguments: '$(BuildName)'
 - publish: artifacts
   artifact: 'yuzu-$(BuildName)-windows-msvc'
   displayName: 'Upload Artifacts'

.gitmodules

@@ -26,11 +26,11 @@
     path = externals/mbedtls
     url = https://github.com/DarkLordZach/mbedtls
 [submodule "opus"]
-    path = externals/opus
-    url = https://github.com/ogniK5377/opus.git
+    path = externals/opus/opus
+    url = https://github.com/xiph/opus.git
 [submodule "soundtouch"]
-	path = externals/soundtouch
-	url = https://github.com/citra-emu/ext-soundtouch.git
+    path = externals/soundtouch
+    url = https://github.com/citra-emu/ext-soundtouch.git
 [submodule "libressl"]
     path = externals/libressl
     url = https://github.com/citra-emu/ext-libressl-portable.git
@@ -47,8 +47,8 @@
     path = externals/sirit
     url = https://github.com/ReinUsesLisp/sirit
 [submodule "libzip"]
-    path = externals/libzip
-    url = https://github.com/DarkLordZach/libzip
+    path = externals/libzip/libzip
+    url = https://github.com/nih-at/libzip.git
 [submodule "zlib"]
-    path = externals/zlib
-    url = https://github.com/madler/zlib
+    path = externals/zlib/zlib
+    url = https://github.com/madler/zlib.git

externals/CMakeLists.txt

@@ -76,6 +76,7 @@ endif()
 
 # zlib
 add_subdirectory(zlib EXCLUDE_FROM_ALL)
+set(ZLIB_LIBRARIES z)
 
 # libzip
 add_subdirectory(libzip EXCLUDE_FROM_ALL)

externals/libzip/CMakeLists.txt

@@ -0,0 +1,564 @@
+# TODO:
+# create usable libtool .la file
+
+CMAKE_MINIMUM_REQUIRED(VERSION 3.0.2)
+
+LIST(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/libzip")
+
+PROJECT(libzip C)
+
+OPTION(ENABLE_COMMONCRYPTO "Enable use of CommonCrypto" ON)
+OPTION(ENABLE_GNUTLS "Enable use of GnuTLS" ON)
+OPTION(ENABLE_MBEDTLS "Enable use of mbed TLS" ON)
+OPTION(ENABLE_OPENSSL "Enable use of OpenSSL" ON)
+OPTION(ENABLE_WINDOWS_CRYPTO "Enable use of Windows cryptography libraries" ON)
+
+OPTION(ENABLE_BZIP2 "Enable use of BZip2" OFF)
+OPTION(ENABLE_LZMA "Enable use of LZMA" OFF)
+
+INCLUDE(CheckFunctionExists)
+INCLUDE(CheckIncludeFiles)
+INCLUDE(CheckSymbolExists)
+INCLUDE(CheckTypeSize)
+INCLUDE(CheckCSourceRuns)
+INCLUDE(CheckCSourceCompiles)
+INCLUDE(CheckStructHasMember)
+INCLUDE(TestBigEndian)
+INCLUDE(GNUInstallDirs)
+IF(ENABLE_COMMONCRYPTO)
+  CHECK_INCLUDE_FILES(CommonCrypto/CommonCrypto.h COMMONCRYPTO_FOUND)
+ELSE()
+  SET(COMMONCRYPTO_FOUND FALSE)
+ENDIF()
+IF(ENABLE_GNUTLS)
+  INCLUDE(FindNettle)
+  INCLUDE(FindGnuTLS)
+ELSE()
+  SET(GNUTLS_FOUND FALSE)
+ENDIF()
+IF(ENABLE_MBEDTLS)
+  FIND_PATH(MBEDTLS_INCLUDE_DIR mbedtls/aes.h)
+  FIND_LIBRARY(MBEDTLS_LIBRARIES NAMES mbedcrypto)
+ELSE()
+  SET(MBEDTLS_LIBRARIES FALSE)
+ENDIF()
+IF(ENABLE_OPENSSL)
+  INCLUDE(FindOpenSSL)
+ELSE()
+  SET(OPENSSL_FOUND FALSE)
+ENDIF()
+IF(WIN32)
+  IF(ENABLE_WINDOWS_CRYPTO)
+    SET(WINDOWS_CRYPTO_FOUND TRUE)
+  ENDIF()
+ELSE()
+  SET(WINDOWS_CRYPTO_FOUND FALSE)
+ENDIF()
+
+OPTION(BUILD_SHARED_LIBS "Build shared libraries" ON)
+OPTION(SHARED_LIB_VERSIONNING "Add SO version in .so build" ON)
+
+SET(PACKAGE "libzip")
+SET(PACKAGE_NAME ${PACKAGE})
+SET(PACKAGE_VERSION_MAJOR "1")
+SET(PACKAGE_VERSION_MINOR "5")
+SET(PACKAGE_VERSION_MICRO "2a")
+#SET(VERSION "${PACKAGE_VERSION_MAJOR}.${PACKAGE_VERSION_MINOR}")
+SET(VERSION "${PACKAGE_VERSION_MAJOR}.${PACKAGE_VERSION_MINOR}.${PACKAGE_VERSION_MICRO}")
+SET(PACKAGE_VERSION ${VERSION})
+SET(LIBZIP_VERSION ${PACKAGE_VERSION})
+SET(LIBZIP_VERSION_MAJOR ${PACKAGE_VERSION_MAJOR})
+SET(LIBZIP_VERSION_MINOR ${PACKAGE_VERSION_MINOR})
+SET(LIBZIP_VERSION_MICRO ${PACKAGE_VERSION_MICRO})
+SET(PACKAGE_STRING "${PACKAGE_NAME} ${PACKAGE_VERSION}")
+
+SET(ARCHIVE_NAME ${PACKAGE_NAME}-${PACKAGE_VERSION})
+IF(NOT TARGET dist)
+ADD_CUSTOM_TARGET(dist
+  COMMAND git config tar.tar.xz.command "xz -c"
+  COMMAND git archive --prefix=${ARCHIVE_NAME}/ -o ${ARCHIVE_NAME}.tar.gz HEAD
+  COMMAND git archive --prefix=${ARCHIVE_NAME}/ -o ${ARCHIVE_NAME}.tar.xz HEAD
+  WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}
+  )
+ADD_CUSTOM_TARGET(distcheck
+  COMMAND chmod -R u+w ${ARCHIVE_NAME} ${ARCHIVE_NAME}-build ${ARCHIVE_NAME}-dest 2>/dev/null || true
+  COMMAND rm -rf ${ARCHIVE_NAME} ${ARCHIVE_NAME}-build ${ARCHIVE_NAME}-dest
+  COMMAND cmake -E tar xf ${ARCHIVE_NAME}.tar.gz
+  COMMAND chmod -R u-w ${ARCHIVE_NAME}
+  COMMAND mkdir ${ARCHIVE_NAME}-build
+  COMMAND mkdir ${ARCHIVE_NAME}-dest
+  COMMAND cd ${ARCHIVE_NAME}-build && cmake -DCMAKE_INSTALL_PREFIX=../${ARCHIVE_NAME}-dest ../${ARCHIVE_NAME}
+  COMMAND cd ${ARCHIVE_NAME}-build && make -j4
+  COMMAND cd ${ARCHIVE_NAME}-build && make test
+  COMMAND cd ${ARCHIVE_NAME}-build && make install
+#  COMMAND cd ${ARCHIVE_NAME}-build && make uninstall
+#  COMMAND if [ `find ${ARCHIVE_NAME}-dest ! -type d | wc -l` -ne 0 ]; then echo leftover files in ${ARCHIVE_NAME}-dest; false; fi
+  COMMAND cd ${ARCHIVE_NAME}-build && make clean
+  COMMAND chmod -R u+w ${ARCHIVE_NAME} ${ARCHIVE_NAME}-build ${ARCHIVE_NAME}-dest
+  COMMAND rm -rf ${ARCHIVE_NAME} ${ARCHIVE_NAME}-build ${ARCHIVE_NAME}-dest
+  COMMAND echo "${ARCHIVE_NAME}.tar.gz is ready for distribution."
+  WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}
+  )
+ADD_DEPENDENCIES(distcheck dist)
+ENDIF(NOT TARGET dist)
+
+IF(BUILD_SHARED_LIBS)
+  SET(HAVE_SHARED TRUE)
+ELSE()
+  SET(ZIP_STATIC TRUE)
+ENDIF()
+
+# Checks
+
+CHECK_FUNCTION_EXISTS(_chmod HAVE__CHMOD)
+CHECK_FUNCTION_EXISTS(_close HAVE__CLOSE)
+CHECK_FUNCTION_EXISTS(_dup HAVE__DUP)
+CHECK_FUNCTION_EXISTS(_fdopen HAVE__FDOPEN)
+CHECK_FUNCTION_EXISTS(_fileno HAVE__FILENO)
+CHECK_FUNCTION_EXISTS(_open HAVE__OPEN)
+CHECK_FUNCTION_EXISTS(_setmode HAVE__SETMODE)
+CHECK_FUNCTION_EXISTS(_snprintf HAVE__SNPRINTF)
+CHECK_FUNCTION_EXISTS(_strdup HAVE__STRDUP)
+CHECK_FUNCTION_EXISTS(_stricmp HAVE__STRICMP)
+CHECK_FUNCTION_EXISTS(_strtoi64 HAVE__STRTOI64)
+CHECK_FUNCTION_EXISTS(_strtoui64 HAVE__STRTOUI64)
+CHECK_FUNCTION_EXISTS(_unlink HAVE__UNLINK)
+CHECK_FUNCTION_EXISTS(arc4random HAVE_ARC4RANDOM)
+CHECK_FUNCTION_EXISTS(clonefile HAVE_CLONEFILE)
+CHECK_FUNCTION_EXISTS(explicit_bzero HAVE_EXPLICIT_BZERO)
+CHECK_FUNCTION_EXISTS(explicit_memset HAVE_EXPLICIT_MEMSET)
+CHECK_FUNCTION_EXISTS(fileno HAVE_FILENO)
+CHECK_FUNCTION_EXISTS(fseeko HAVE_FSEEKO)
+CHECK_FUNCTION_EXISTS(ftello HAVE_FTELLO)
+CHECK_FUNCTION_EXISTS(getprogname HAVE_GETPROGNAME)
+CHECK_FUNCTION_EXISTS(localtime_r HAVE_LOCALTIME_R)
+CHECK_FUNCTION_EXISTS(open HAVE_OPEN)
+CHECK_FUNCTION_EXISTS(setmode HAVE_SETMODE)
+CHECK_FUNCTION_EXISTS(snprintf HAVE_SNPRINTF)
+CHECK_FUNCTION_EXISTS(strcasecmp HAVE_STRCASECMP)
+CHECK_FUNCTION_EXISTS(strdup HAVE_STRDUP)
+CHECK_FUNCTION_EXISTS(stricmp HAVE_STRICMP)
+CHECK_FUNCTION_EXISTS(strtoll HAVE_STRTOLL)
+CHECK_FUNCTION_EXISTS(strtoull HAVE_STRTOULL)
+
+CHECK_INCLUDE_FILES("sys/types.h;sys/stat.h;fts.h" HAVE_FTS_H)
+CHECK_INCLUDE_FILES(stdbool.h HAVE_STDBOOL_H)
+CHECK_INCLUDE_FILES(strings.h HAVE_STRINGS_H)
+CHECK_INCLUDE_FILES(unistd.h HAVE_UNISTD_H)
+
+CHECK_INCLUDE_FILES(inttypes.h HAVE_INTTYPES_H_LIBZIP)
+CHECK_INCLUDE_FILES(stdint.h HAVE_STDINT_H_LIBZIP)
+CHECK_INCLUDE_FILES(sys/types.h HAVE_SYS_TYPES_H_LIBZIP)
+
+# TODO: fix test
+# this test does not find __progname even when it exists
+#CHECK_SYMBOL_EXISTS(__progname stdlib.h HAVE___PROGNAME)
+
+CHECK_TYPE_SIZE(__int8 __INT8_LIBZIP)
+CHECK_TYPE_SIZE(int8_t INT8_T_LIBZIP)
+CHECK_TYPE_SIZE(uint8_t UINT8_T_LIBZIP)
+CHECK_TYPE_SIZE(__int16 __INT16_LIBZIP)
+CHECK_TYPE_SIZE(int16_t INT16_T_LIBZIP)
+CHECK_TYPE_SIZE(uint16_t UINT16_T_LIBZIP)
+CHECK_TYPE_SIZE(__int32 __INT32_LIBZIP)
+CHECK_TYPE_SIZE(int32_t INT32_T_LIBZIP)
+CHECK_TYPE_SIZE(uint32_t UINT32_T_LIBZIP)
+CHECK_TYPE_SIZE(__int64 __INT64_LIBZIP)
+CHECK_TYPE_SIZE(int64_t INT64_T_LIBZIP)
+CHECK_TYPE_SIZE(uint64_t UINT64_T_LIBZIP)
+CHECK_TYPE_SIZE("short" SHORT_LIBZIP)
+CHECK_TYPE_SIZE("int" INT_LIBZIP)
+CHECK_TYPE_SIZE("long" LONG_LIBZIP)
+CHECK_TYPE_SIZE("long long" LONG_LONG_LIBZIP)
+CHECK_TYPE_SIZE("off_t" SIZEOF_OFF_T)
+CHECK_TYPE_SIZE("size_t" SIZE_T_LIBZIP)
+CHECK_TYPE_SIZE("ssize_t" SSIZE_T_LIBZIP)
+
+CHECK_C_SOURCE_COMPILES("#include <sys/ioctl.h>
+#include <linux/fs.h>
+int main(int argc, char *argv[]) { unsigned long x = FICLONERANGE; }" HAVE_FICLONERANGE)
+
+CHECK_C_SOURCE_COMPILES("
+int foo(char * _Nullable bar);
+int main(int argc, char *argv[]) { }" HAVE_NULLABLE)
+
+TEST_BIG_ENDIAN(WORDS_BIGENDIAN)
+
+#FIND_PACKAGE(ZLIB 1.1.2 REQUIRED)
+INCLUDE_DIRECTORIES(../zlib/zlib)
+SET(CMAKE_REQUIRED_INCLUDES ../zlib/zlib)
+
+IF(ENABLE_BZIP2)
+  FIND_PACKAGE(BZip2)
+  IF(BZIP2_FOUND)
+    SET (HAVE_LIBBZ2 1)
+
+    INCLUDE_DIRECTORIES(${BZIP2_INCLUDE_DIR})
+    SET (OPTIONAL_LIBRARY ${OPTIONAL_LIBRARY} ${BZIP2_LIBRARIES})
+  ELSE()
+    MESSAGE(WARNING "-- bzip2 library not found; bzip2 support disabled")
+  ENDIF(BZIP2_FOUND)
+ENDIF(ENABLE_BZIP2)
+
+IF(ENABLE_LZMA)
+  FIND_PACKAGE(LibLZMA)
+  IF(LIBLZMA_FOUND)
+    SET (HAVE_LIBLZMA 1)
+
+    INCLUDE_DIRECTORIES(${LIBLZMA_INCLUDE_DIR})
+    SET (OPTIONAL_LIBRARY ${OPTIONAL_LIBRARY} ${LIBLZMA_LIBRARY})
+  ELSE()
+    MESSAGE(WARNING "-- lzma library not found; lzma support disabled")
+  ENDIF(LIBLZMA_FOUND)
+ENDIF(ENABLE_LZMA)
+
+
+IF (COMMONCRYPTO_FOUND)
+  SET (HAVE_CRYPTO 1)
+  SET (HAVE_COMMONCRYPTO 1)
+ELSEIF (WINDOWS_CRYPTO_FOUND)
+  SET (HAVE_CRYPTO 1)
+  SET (HAVE_WINDOWS_CRYPTO 1)
+ELSEIF (GNUTLS_FOUND AND NETTLE_FOUND)
+  SET (HAVE_CRYPTO 1)
+  SET (HAVE_GNUTLS 1)
+  INCLUDE_DIRECTORIES(${GNUTLS_INCLUDE_DIR} ${NETTLE_INCLUDE_DIR})
+  SET (OPTIONAL_LIBRARY ${OPTIONAL_LIBRARY} ${GNUTLS_LIBRARY} ${NETTLE_LIBRARY})
+ELSEIF (OPENSSL_FOUND)
+  SET (HAVE_CRYPTO 1)
+  SET (HAVE_OPENSSL 1)
+  INCLUDE_DIRECTORIES(${OPENSSL_INCLUDE_DIR})
+  SET (OPTIONAL_LIBRARY ${OPTIONAL_LIBRARY} ${OPENSSL_LIBRARIES})
+ELSEIF (MBEDTLS_LIBRARIES)
+  SET (HAVE_CRYPTO 1)
+  SET (HAVE_MBEDTLS 1)
+  INCLUDE_DIRECTORIES(${MBEDTLS_INCLUDE_DIR})
+  SET (OPTIONAL_LIBRARY ${OPTIONAL_LIBRARY} ${MBEDTLS_LIBRARIES})
+ENDIF()
+
+IF (NOT HAVE_CRYPTO)
+  MESSAGE(WARNING "-- neither Common Crypto, GnuTLS, mbed TLS, OpenSSL, nor Windows Cryptography found; AES support disabled")
+ENDIF()
+
+IF(MSVC)
+ADD_DEFINITIONS("-D_CRT_SECURE_NO_WARNINGS")
+ADD_DEFINITIONS("-D_CRT_NONSTDC_NO_DEPRECATE")
+ENDIF(MSVC)
+
+if(WIN32)
+  if(HAVE_WINDOWS_CRYPTO)
+    SET (OPTIONAL_LIBRARY ${OPTIONAL_LIBRARY} bcrypt)
+  endif()
+  if(CMAKE_SYSTEM_NAME MATCHES WindowsPhone OR CMAKE_SYSTEM_NAME MATCHES WindowsStore)
+    ADD_DEFINITIONS(-DMS_UWP)
+  else(CMAKE_SYSTEM_NAME MATCHES WindowsPhone OR CMAKE_SYSTEM_NAME MATCHES WindowsStore)
+    SET (OPTIONAL_LIBRARY ${OPTIONAL_LIBRARY} advapi32)
+  endif(CMAKE_SYSTEM_NAME MATCHES WindowsPhone OR CMAKE_SYSTEM_NAME MATCHES WindowsStore)
+endif(WIN32)
+
+ADD_DEFINITIONS("-DHAVE_CONFIG_H")
+
+# rpath handling: use rpath in installed binaries
+IF(NOT CMAKE_SYSTEM_NAME MATCHES Linux)
+	SET(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR}")
+	SET(CMAKE_INSTALL_RPATH_USE_LINK_PATH TRUE)
+ENDIF()
+
+# fixed size integral types
+
+IF(HAVE_INTTYPES_H_LIBZIP)
+  SET(LIBZIP_TYPES_INCLUDE "#define __STDC_FORMAT_MACROS 1
+#include <inttypes.h>")
+ELSEIF(HAVE_STDINT_H_LIBZIP)
+  SET(LIBZIP_TYPES_INCLUDE "#include <stdint.h>")
+ELSEIF(HAVE_SYS_TYPES_H_LIBZIP)
+  SET(LIBZIP_TYPES_INCLUDE "#include <sys/types.h>")
+ENDIF()
+
+IF(HAVE_INT8_T_LIBZIP)
+  SET(ZIP_INT8_T int8_t)
+ELSEIF(HAVE___INT8_LIBZIP)
+  SET(ZIP_INT8_T __int8)
+ELSE()
+  SET(ZIP_INT8_T "signed char")
+ENDIF()
+
+IF(HAVE_UINT8_T_LIBZIP)
+  SET(ZIP_UINT8_T uint8_t)
+ELSEIF(HAVE___INT8_LIBZIP)
+  SET(ZIP_UINT8_T "unsigned __int8")
+ELSE()
+  SET(ZIP_UINT8_T "unsigned char")
+ENDIF()
+
+IF(HAVE_INT16_T_LIBZIP)
+  SET(ZIP_INT16_T int16_t)
+ELSEIF(HAVE___INT16_LIBZIP)
+  SET(INT16_T_LIBZIP __int16)
+ELSEIF(SHORT_LIBZIP EQUAL 2)
+  SET(INT16_T_LIBZIP short)
+ENDIF()
+
+IF(HAVE_UINT16_T_LIBZIP)
+  SET(ZIP_UINT16_T uint16_t)
+ELSEIF(HAVE___INT16_LIBZIP)
+  SET(UINT16_T_LIBZIP "unsigned __int16")
+ELSEIF(SHORT_LIBZIP EQUAL 2)
+  SET(UINT16_T_LIBZIP "unsigned short")
+ENDIF()
+
+IF(HAVE_INT32_T_LIBZIP)
+  SET(ZIP_INT32_T int32_t)
+ELSEIF(HAVE___INT32_LIBZIP)
+  SET(ZIP_INT32_T __int32)
+ELSEIF(INT_LIBZIP EQUAL 4)
+  SET(ZIP_INT32_T int)
+ELSEIF(LONG_LIBZIP EQUAL 4)
+  SET(ZIP_INT32_T long)
+ENDIF()
+
+IF(HAVE_UINT32_T_LIBZIP)
+SET(ZIP_UINT32_T uint32_t)
+ELSEIF(HAVE___INT32_LIBZIP)
+SET(ZIP_UINT32_T "unsigned __int32")
+ELSEIF(INT_LIBZIP EQUAL 4)
+SET(ZIP_UINT32_T "unsigned int")
+ELSEIF(LONG_LIBZIP EQUAL 4)
+SET(ZIP_UINT32_T "unsigned long")
+ENDIF()
+
+IF(HAVE_INT64_T_LIBZIP)
+  SET(ZIP_INT64_T int64_t)
+ELSEIF(HAVE___INT64_LIBZIP)
+  SET(ZIP_INT64_T __int64)
+ELSEIF(LONG_LIBZIP EQUAL 8)
+  SET(ZIP_INT64_T long)
+ELSEIF(LONG_LONG_LIBZIP EQUAL 8)
+  SET(ZIP_INT64_T "long long")
+ENDIF()
+
+IF(HAVE_UINT64_T_LIBZIP)
+  SET(ZIP_UINT64_T uint64_t)
+ELSEIF(HAVE___INT64_LIBZIP)
+  SET(ZIP_UINT64_T "unsigned __int64")
+ELSEIF(LONG_LIBZIP EQUAL 8)
+  SET(ZIP_UINT64_T "unsigned long")
+ELSEIF(LONG_LONG_LIBZIP EQUAL 8)
+  SET(ZIP_UINT64_T "unsigned long long")
+ENDIF()
+
+IF(HAVE_NULLABLE)
+  SET(ZIP_NULLABLE_DEFINES)
+ELSE()
+  SET(ZIP_NULLABLE_DEFINES "#define _Nullable
+#define _Nonnull")
+ENDIF()
+
+# write out config file
+CONFIGURE_FILE(${CMAKE_CURRENT_SOURCE_DIR}/libzip/cmake-config.h.in ${CMAKE_CURRENT_BINARY_DIR}/libzip/config.h)
+CONFIGURE_FILE(${CMAKE_CURRENT_SOURCE_DIR}/libzip/cmake-zipconf.h.in ${CMAKE_CURRENT_BINARY_DIR}/libzip/zipconf.h)
+
+# installation
+INSTALL(FILES ${CMAKE_CURRENT_BINARY_DIR}/libzip/zipconf.h DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
+INSTALL(FILES libzip/lib/zip.h DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
+
+SET(CMAKE_C_VISIBILITY_PRESET hidden)
+
+ADD_LIBRARY(zip
+  libzip/lib/zip_add.c
+  libzip/lib/zip_add_dir.c
+  libzip/lib/zip_add_entry.c
+  libzip/lib/zip_algorithm_deflate.c
+  libzip/lib/zip_buffer.c
+  libzip/lib/zip_close.c
+  libzip/lib/zip_delete.c
+  libzip/lib/zip_dir_add.c
+  libzip/lib/zip_dirent.c
+  libzip/lib/zip_discard.c
+  libzip/lib/zip_entry.c
+  libzip/lib/zip_err_str.c
+  libzip/lib/zip_error.c
+  libzip/lib/zip_error_clear.c
+  libzip/lib/zip_error_get.c
+  libzip/lib/zip_error_get_sys_type.c
+  libzip/lib/zip_error_strerror.c
+  libzip/lib/zip_error_to_str.c
+  libzip/lib/zip_extra_field.c
+  libzip/lib/zip_extra_field_api.c
+  libzip/lib/zip_fclose.c
+  libzip/lib/zip_fdopen.c
+  libzip/lib/zip_file_add.c
+  libzip/lib/zip_file_error_clear.c
+  libzip/lib/zip_file_error_get.c
+  libzip/lib/zip_file_get_comment.c
+  libzip/lib/zip_file_get_external_attributes.c
+  libzip/lib/zip_file_get_offset.c
+  libzip/lib/zip_file_rename.c
+  libzip/lib/zip_file_replace.c
+  libzip/lib/zip_file_set_comment.c
+  libzip/lib/zip_file_set_encryption.c
+  libzip/lib/zip_file_set_external_attributes.c
+  libzip/lib/zip_file_set_mtime.c
+  libzip/lib/zip_file_strerror.c
+  libzip/lib/zip_filerange_crc.c
+  libzip/lib/zip_fopen.c
+  libzip/lib/zip_fopen_encrypted.c
+  libzip/lib/zip_fopen_index.c
+  libzip/lib/zip_fopen_index_encrypted.c
+  libzip/lib/zip_fread.c
+  libzip/lib/zip_fseek.c
+  libzip/lib/zip_ftell.c
+  libzip/lib/zip_get_archive_comment.c
+  libzip/lib/zip_get_archive_flag.c
+  libzip/lib/zip_get_encryption_implementation.c
+  libzip/lib/zip_get_file_comment.c
+  libzip/lib/zip_get_name.c
+  libzip/lib/zip_get_num_entries.c
+  libzip/lib/zip_get_num_files.c
+  libzip/lib/zip_hash.c
+  libzip/lib/zip_io_util.c
+  libzip/lib/zip_libzip_version.c
+  libzip/lib/zip_memdup.c
+  libzip/lib/zip_name_locate.c
+  libzip/lib/zip_new.c
+  libzip/lib/zip_open.c
+  libzip/lib/zip_progress.c
+  libzip/lib/zip_rename.c
+  libzip/lib/zip_replace.c
+  libzip/lib/zip_set_archive_comment.c
+  libzip/lib/zip_set_archive_flag.c
+  libzip/lib/zip_set_default_password.c
+  libzip/lib/zip_set_file_comment.c
+  libzip/lib/zip_set_file_compression.c
+  libzip/lib/zip_set_name.c
+  libzip/lib/zip_source_accept_empty.c
+  libzip/lib/zip_source_begin_write.c
+  libzip/lib/zip_source_begin_write_cloning.c
+  libzip/lib/zip_source_buffer.c
+  libzip/lib/zip_source_call.c
+  libzip/lib/zip_source_close.c
+  libzip/lib/zip_source_commit_write.c
+  libzip/lib/zip_source_compress.c
+  libzip/lib/zip_source_crc.c
+  libzip/lib/zip_source_error.c
+  libzip/lib/zip_source_filep.c
+  libzip/lib/zip_source_free.c
+  libzip/lib/zip_source_function.c
+  libzip/lib/zip_source_get_compression_flags.c
+  libzip/lib/zip_source_is_deleted.c
+  libzip/lib/zip_source_layered.c
+  libzip/lib/zip_source_open.c
+  libzip/lib/zip_source_pkware.c
+  libzip/lib/zip_source_read.c
+  libzip/lib/zip_source_remove.c
+  libzip/lib/zip_source_rollback_write.c
+  libzip/lib/zip_source_seek.c
+  libzip/lib/zip_source_seek_write.c
+  libzip/lib/zip_source_stat.c
+  libzip/lib/zip_source_supports.c
+  libzip/lib/zip_source_tell.c
+  libzip/lib/zip_source_tell_write.c
+  libzip/lib/zip_source_window.c
+  libzip/lib/zip_source_write.c
+  libzip/lib/zip_source_zip.c
+  libzip/lib/zip_source_zip_new.c
+  libzip/lib/zip_stat.c
+  libzip/lib/zip_stat_index.c
+  libzip/lib/zip_stat_init.c
+  libzip/lib/zip_strerror.c
+  libzip/lib/zip_string.c
+  libzip/lib/zip_unchange.c
+  libzip/lib/zip_unchange_all.c
+  libzip/lib/zip_unchange_archive.c
+  libzip/lib/zip_unchange_data.c
+  libzip/lib/zip_utf-8.c
+)
+
+IF(WIN32)
+  target_sources(zip PRIVATE
+    libzip/lib/zip_source_win32handle.c
+    libzip/lib/zip_source_win32utf8.c
+    libzip/lib/zip_source_win32w.c
+  )
+  IF(CMAKE_SYSTEM_NAME MATCHES WindowsPhone OR CMAKE_SYSTEM_NAME MATCHES WindowsStore)
+  ELSE()
+    target_sources(zip PRIVATE libzip/lib/zip_source_win32a.c)
+  ENDIF()
+ELSE()
+  target_sources(zip PRIVATE
+    libzip/lib/zip_mkstempm.c
+    libzip/lib/zip_source_file.c
+    libzip/lib/zip_random_unix.c
+  )
+ENDIF()
+
+IF(HAVE_LIBBZ2)
+  target_sources(zip PRIVATE libzip/lib/zip_algorithm_bzip2.c)
+ENDIF()
+
+IF(HAVE_LIBLZMA)
+  target_sources(zip PRIVATE libzip/lib/zip_algorithm_xz.c)
+ENDIF()
+
+IF(HAVE_COMMONCRYPTO)
+  target_sources(zip PRIVATE libzip/lib/zip_crypto_commoncrypto.c)
+ELSEIF(HAVE_WINDOWS_CRYPTO)
+  target_sources(zip PRIVATE libzip/lib/zip_crypto_win.c)
+ELSEIF(HAVE_GNUTLS)
+  target_sources(zip PRIVATE libzip/lib/zip_crypto_gnutls.c)
+ELSEIF(HAVE_OPENSSL)
+  target_sources(zip PRIVATE libzip/lib/zip_crypto_openssl.c)
+ELSEIF(HAVE_MBEDTLS)
+  target_sources(zip PRIVATE libzip/lib/zip_crypto_mbedtls.c)
+ENDIF()
+
+IF(HAVE_CRYPTO)
+  target_sources(zip PRIVATE
+    libzip/lib/zip_winzip_aes.c
+    libzip/lib/zip_source_winzip_aes_decode.c
+    libzip/lib/zip_source_winzip_aes_encode.c
+  )
+ENDIF()
+
+target_include_directories(zip
+PUBLIC
+  libzip/lib
+  ${CMAKE_CURRENT_BINARY_DIR}/libzip
+)
+
+# pkgconfig file
+SET(prefix ${CMAKE_INSTALL_PREFIX})
+SET(exec_prefix \${prefix})
+SET(bindir \${exec_prefix}/${CMAKE_INSTALL_BINDIR})
+SET(libdir \${exec_prefix}/${CMAKE_INSTALL_LIBDIR})
+SET(includedir \${prefix}/${CMAKE_INSTALL_INCLUDEDIR})
+IF(CMAKE_SYSTEM_NAME MATCHES BSD)
+  SET(PKG_CONFIG_RPATH "-Wl,-R\${libdir}")
+ENDIF(CMAKE_SYSTEM_NAME MATCHES BSD)
+get_target_property(LIBS_PRIVATE zip LINK_LIBRARIES)
+foreach(LIB ${LIBS_PRIVATE})
+  if(LIB MATCHES "^/")
+    get_filename_component(LIB ${LIB} NAME_WE)
+    string(REGEX REPLACE "^lib" "" LIB ${LIB})
+  endif()
+  set(LIBS "${LIBS} -l${LIB}")
+endforeach()
+CONFIGURE_FILE(libzip/libzip.pc.in libzip/libzip.pc @ONLY)
+INSTALL(FILES ${CMAKE_CURRENT_BINARY_DIR}/libzip.pc DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig)
+
+ADD_CUSTOM_TARGET(update_zip_err_str
+  COMMAND sh ${CMAKE_CURRENT_SOURCE_DIR}/libzip/lib/make_zip_err_str.sh ${CMAKE_CURRENT_SOURCE_DIR}/libzip/lib/zip.h ${CMAKE_CURRENT_SOURCE_DIR}/libzip/lib/zip_err_str.c
+  DEPENDS ${CMAKE_CURRENT_SOURCE_DIR}/libzip/lib/zip.h ${CMAKE_CURRENT_SOURCE_DIR}/libzip/lib/make_zip_err_str.sh
+)
+
+IF(SHARED_LIB_VERSIONNING)
+SET_TARGET_PROPERTIES(zip PROPERTIES VERSION 5.0 SOVERSION 5)
+ENDIF()
+
+TARGET_LINK_LIBRARIES(zip ${ZLIB_LIBRARIES} ${OPTIONAL_LIBRARY})
+INSTALL(TARGETS zip
+  RUNTIME DESTINATION ${CMAKE_INSTALL_BINDIR}
+  ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR}
+  LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR}
+)
+

externals/opus/CMakeLists.txt

@@ -0,0 +1,250 @@
+cmake_minimum_required(VERSION 3.8)
+
+project(opus)
+
+option(OPUS_STACK_PROTECTOR "Use stack protection" OFF)
+option(OPUS_USE_ALLOCA "Use alloca for stack arrays (on non-C99 compilers)" OFF)
+option(OPUS_CUSTOM_MODES "Enable non-Opus modes, e.g. 44.1 kHz & 2^n frames" OFF)
+option(OPUS_FIXED_POINT "Compile as fixed-point (for machines without a fast enough FPU)" OFF)
+option(OPUS_ENABLE_FLOAT_API "Compile with the floating point API (for machines with float library" ON)
+
+include(opus/opus_functions.cmake)
+
+if(OPUS_STACK_PROTECTOR)
+    if(NOT MSVC) # GC on by default on MSVC
+        check_and_set_flag(STACK_PROTECTION_STRONG -fstack-protector-strong)
+    endif()
+else()
+    if(MSVC)
+        check_and_set_flag(BUFFER_SECURITY_CHECK /GS-)
+    endif()
+endif()
+
+add_library(opus STATIC
+    # CELT sources
+    opus/celt/bands.c
+    opus/celt/celt.c
+    opus/celt/celt_decoder.c
+    opus/celt/celt_encoder.c
+    opus/celt/celt_lpc.c
+    opus/celt/cwrs.c
+    opus/celt/entcode.c
+    opus/celt/entdec.c
+    opus/celt/entenc.c
+    opus/celt/kiss_fft.c
+    opus/celt/laplace.c
+    opus/celt/mathops.c
+    opus/celt/mdct.c
+    opus/celt/modes.c
+    opus/celt/pitch.c
+    opus/celt/quant_bands.c
+    opus/celt/rate.c
+    opus/celt/vq.c
+
+    # SILK sources
+    opus/silk/A2NLSF.c
+    opus/silk/CNG.c
+    opus/silk/HP_variable_cutoff.c
+    opus/silk/LPC_analysis_filter.c
+    opus/silk/LPC_fit.c
+    opus/silk/LPC_inv_pred_gain.c
+    opus/silk/LP_variable_cutoff.c
+    opus/silk/NLSF2A.c
+    opus/silk/NLSF_VQ.c
+    opus/silk/NLSF_VQ_weights_laroia.c
+    opus/silk/NLSF_decode.c
+    opus/silk/NLSF_del_dec_quant.c
+    opus/silk/NLSF_encode.c
+    opus/silk/NLSF_stabilize.c
+    opus/silk/NLSF_unpack.c
+    opus/silk/NSQ.c
+    opus/silk/NSQ_del_dec.c
+    opus/silk/PLC.c
+    opus/silk/VAD.c
+    opus/silk/VQ_WMat_EC.c
+    opus/silk/ana_filt_bank_1.c
+    opus/silk/biquad_alt.c
+    opus/silk/bwexpander.c
+    opus/silk/bwexpander_32.c
+    opus/silk/check_control_input.c
+    opus/silk/code_signs.c
+    opus/silk/control_SNR.c
+    opus/silk/control_audio_bandwidth.c
+    opus/silk/control_codec.c
+    opus/silk/dec_API.c
+    opus/silk/decode_core.c
+    opus/silk/decode_frame.c
+    opus/silk/decode_indices.c
+    opus/silk/decode_parameters.c
+    opus/silk/decode_pitch.c
+    opus/silk/decode_pulses.c
+    opus/silk/decoder_set_fs.c
+    opus/silk/enc_API.c
+    opus/silk/encode_indices.c
+    opus/silk/encode_pulses.c
+    opus/silk/gain_quant.c
+    opus/silk/init_decoder.c
+    opus/silk/init_encoder.c
+    opus/silk/inner_prod_aligned.c
+    opus/silk/interpolate.c
+    opus/silk/lin2log.c
+    opus/silk/log2lin.c
+    opus/silk/pitch_est_tables.c
+    opus/silk/process_NLSFs.c
+    opus/silk/quant_LTP_gains.c
+    opus/silk/resampler.c
+    opus/silk/resampler_down2.c
+    opus/silk/resampler_down2_3.c
+    opus/silk/resampler_private_AR2.c
+    opus/silk/resampler_private_IIR_FIR.c
+    opus/silk/resampler_private_down_FIR.c
+    opus/silk/resampler_private_up2_HQ.c
+    opus/silk/resampler_rom.c
+    opus/silk/shell_coder.c
+    opus/silk/sigm_Q15.c
+    opus/silk/sort.c
+    opus/silk/stereo_LR_to_MS.c
+    opus/silk/stereo_MS_to_LR.c
+    opus/silk/stereo_decode_pred.c
+    opus/silk/stereo_encode_pred.c
+    opus/silk/stereo_find_predictor.c
+    opus/silk/stereo_quant_pred.c
+    opus/silk/sum_sqr_shift.c
+    opus/silk/table_LSF_cos.c
+    opus/silk/tables_LTP.c
+    opus/silk/tables_NLSF_CB_NB_MB.c
+    opus/silk/tables_NLSF_CB_WB.c
+    opus/silk/tables_gain.c
+    opus/silk/tables_other.c
+    opus/silk/tables_pitch_lag.c
+    opus/silk/tables_pulses_per_block.c
+
+    # Opus sources
+    opus/src/analysis.c
+    opus/src/mapping_matrix.c
+    opus/src/mlp.c
+    opus/src/mlp_data.c
+    opus/src/opus.c
+    opus/src/opus_decoder.c
+    opus/src/opus_encoder.c
+    opus/src/opus_multistream.c
+    opus/src/opus_multistream_decoder.c
+    opus/src/opus_multistream_encoder.c
+    opus/src/opus_projection_decoder.c
+    opus/src/opus_projection_encoder.c
+    opus/src/repacketizer.c
+)
+
+if (DEBUG)
+    target_sources(opus PRIVATE opus/silk/debug.c)
+endif()
+
+if (OPUS_FIXED_POINT)
+    target_sources(opus PRIVATE
+        opus/silk/fixed/LTP_analysis_filter_FIX.c
+        opus/silk/fixed/LTP_scale_ctrl_FIX.c
+        opus/silk/fixed/apply_sine_window_FIX.c
+        opus/silk/fixed/autocorr_FIX.c
+        opus/silk/fixed/burg_modified_FIX.c
+        opus/silk/fixed/corrMatrix_FIX.c
+        opus/silk/fixed/encode_frame_FIX.c
+        opus/silk/fixed/find_LPC_FIX.c
+        opus/silk/fixed/find_LTP_FIX.c
+        opus/silk/fixed/find_pitch_lags_FIX.c
+        opus/silk/fixed/find_pred_coefs_FIX.c
+        opus/silk/fixed/k2a_FIX.c
+        opus/silk/fixed/k2a_Q16_FIX.c
+        opus/silk/fixed/noise_shape_analysis_FIX.c
+        opus/silk/fixed/pitch_analysis_core_FIX.c
+        opus/silk/fixed/prefilter_FIX.c
+        opus/silk/fixed/process_gains_FIX.c
+        opus/silk/fixed/regularize_correlations_FIX.c
+        opus/silk/fixed/residual_energy16_FIX.c
+        opus/silk/fixed/residual_energy_FIX.c
+        opus/silk/fixed/schur64_FIX.c
+        opus/silk/fixed/schur_FIX.c
+        opus/silk/fixed/solve_LS_FIX.c
+        opus/silk/fixed/vector_ops_FIX.c
+        opus/silk/fixed/warped_autocorrelation_FIX.c
+    )
+else()
+    target_sources(opus PRIVATE
+        opus/silk/float/LPC_analysis_filter_FLP.c
+        opus/silk/float/LPC_inv_pred_gain_FLP.c
+        opus/silk/float/LTP_analysis_filter_FLP.c
+        opus/silk/float/LTP_scale_ctrl_FLP.c
+        opus/silk/float/apply_sine_window_FLP.c
+        opus/silk/float/autocorrelation_FLP.c
+        opus/silk/float/burg_modified_FLP.c
+        opus/silk/float/bwexpander_FLP.c
+        opus/silk/float/corrMatrix_FLP.c
+        opus/silk/float/encode_frame_FLP.c
+        opus/silk/float/energy_FLP.c
+        opus/silk/float/find_LPC_FLP.c
+        opus/silk/float/find_LTP_FLP.c
+        opus/silk/float/find_pitch_lags_FLP.c
+        opus/silk/float/find_pred_coefs_FLP.c
+        opus/silk/float/inner_product_FLP.c
+        opus/silk/float/k2a_FLP.c
+        opus/silk/float/noise_shape_analysis_FLP.c
+        opus/silk/float/pitch_analysis_core_FLP.c
+        opus/silk/float/process_gains_FLP.c
+        opus/silk/float/regularize_correlations_FLP.c
+        opus/silk/float/residual_energy_FLP.c
+        opus/silk/float/scale_copy_vector_FLP.c
+        opus/silk/float/scale_vector_FLP.c
+        opus/silk/float/schur_FLP.c
+        opus/silk/float/sort_FLP.c
+        opus/silk/float/warped_autocorrelation_FLP.c
+        opus/silk/float/wrappers_FLP.c
+    )
+endif()
+
+target_compile_definitions(opus PRIVATE OPUS_BUILD ENABLE_HARDENING)
+
+if(NOT MSVC)
+    target_compile_definitions(opus PRIVATE _FORTIFY_SOURCE=2)
+endif()
+
+# It is strongly recommended to uncomment one of these VAR_ARRAYS: Use C99
+# variable-length arrays for stack allocation USE_ALLOCA: Use alloca() for stack
+# allocation If none is defined, then the fallback is a non-threadsafe global
+# array
+if(OPUS_USE_ALLOCA OR MSVC)
+    target_compile_definitions(opus PRIVATE USE_ALLOCA)
+else()
+    target_compile_definitions(opus PRIVATE VAR_ARRAYS)
+endif()
+
+if(OPUS_CUSTOM_MODES)
+    target_compile_definitions(opus PRIVATE CUSTOM_MODES)
+endif()
+
+if(NOT OPUS_ENABLE_FLOAT_API)
+    target_compile_definitions(opus PRIVATE DISABLE_FLOAT_API)
+endif()
+
+target_compile_definitions(opus
+PUBLIC
+    -DOPUS_VERSION="\\"1.3.1\\""
+
+PRIVATE
+    # Use C99 intrinsics to speed up float-to-int conversion
+    HAVE_LRINTF
+)
+
+if (FIXED_POINT)
+    target_compile_definitions(opus PRIVATE -DFIXED_POINT=1 -DDISABLE_FLOAT_API)
+endif()
+
+target_include_directories(opus
+PUBLIC
+    opus/include
+
+PRIVATE
+    opus/celt
+    opus/silk
+    opus/silk/fixed
+    opus/silk/float
+    opus/src
+)

externals/zlib/CMakeLists.txt

@@ -0,0 +1,81 @@
+project(zlib C)
+
+include(CheckTypeSize)
+include(CheckFunctionExists)
+include(CheckIncludeFile)
+
+check_include_file(sys/types.h HAVE_SYS_TYPES_H)
+check_include_file(stdint.h    HAVE_STDINT_H)
+check_include_file(stddef.h    HAVE_STDDEF_H)
+
+# Check to see if we have large file support
+set(CMAKE_REQUIRED_DEFINITIONS -D_LARGEFILE64_SOURCE=1)
+# We add these other definitions here because CheckTypeSize.cmake
+# in CMake 2.4.x does not automatically do so and we want
+# compatibility with CMake 2.4.x.
+if(HAVE_SYS_TYPES_H)
+    list(APPEND CMAKE_REQUIRED_DEFINITIONS -DHAVE_SYS_TYPES_H)
+endif()
+if(HAVE_STDINT_H)
+    list(APPEND CMAKE_REQUIRED_DEFINITIONS -DHAVE_STDINT_H)
+endif()
+if(HAVE_STDDEF_H)
+    list(APPEND CMAKE_REQUIRED_DEFINITIONS -DHAVE_STDDEF_H)
+endif()
+check_type_size(off64_t OFF64_T)
+if(HAVE_OFF64_T)
+   add_definitions(-D_LARGEFILE64_SOURCE=1)
+endif()
+set(CMAKE_REQUIRED_DEFINITIONS) # clear variable
+
+# Check for fseeko
+check_function_exists(fseeko HAVE_FSEEKO)
+if(NOT HAVE_FSEEKO)
+    add_definitions(-DNO_FSEEKO)
+endif()
+
+# Check for unistd.h
+check_include_file(unistd.h HAVE_UNISTD_H)
+if(HAVE_UNISTD_H)
+    add_definitions(-DHAVE_UNISTD_H)
+endif()
+
+if(MSVC)
+    add_definitions(-D_CRT_SECURE_NO_DEPRECATE)
+    add_definitions(-D_CRT_NONSTDC_NO_DEPRECATE)
+endif()
+
+add_library(z STATIC
+    zlib/adler32.c
+    zlib/compress.c
+    zlib/crc32.c
+    zlib/crc32.h
+    zlib/deflate.c
+    zlib/deflate.h
+    zlib/gzclose.c
+    zlib/gzguts.h
+    zlib/gzlib.c
+    zlib/gzread.c
+    zlib/gzwrite.c
+    zlib/inffast.h
+    zlib/inffixed.h
+    zlib/inflate.c
+    zlib/inflate.h
+    zlib/infback.c
+    zlib/inftrees.c
+    zlib/inftrees.h
+    zlib/inffast.c
+    zlib/trees.c
+    zlib/trees.h
+    zlib/uncompr.c
+    zlib/zconf.h
+    zlib/zlib.h
+    zlib/zutil.c
+    zlib/zutil.h
+)
+add_library(ZLIB::ZLIB ALIAS z)
+
+target_include_directories(z
+PUBLIC
+    zlib/
+)

src/common/CMakeLists.txt

@@ -3,17 +3,8 @@
 # could affect the result, but much more unlikely than the following files. Keeping a list of files
 # like this allows for much better caching since it doesn't force the user to recompile binary shaders every update
 set(VIDEO_CORE "${CMAKE_SOURCE_DIR}/src/video_core")
-if (DEFINED ENV{CI})
-  if (DEFINED ENV{TRAVIS})
-    set(BUILD_REPOSITORY $ENV{TRAVIS_REPO_SLUG})
-    set(BUILD_TAG $ENV{TRAVIS_TAG})
-  elseif(DEFINED ENV{APPVEYOR})
-    set(BUILD_REPOSITORY $ENV{APPVEYOR_REPO_NAME})
-    set(BUILD_TAG $ENV{APPVEYOR_REPO_TAG_NAME})
-  elseif(DEFINED ENV{AZURE})
-    set(BUILD_REPOSITORY $ENV{AZURE_REPO_NAME})
-    set(BUILD_TAG $ENV{AZURE_REPO_TAG})
-  endif()
+if (DEFINED ENV{AZURECIREPO})
+  set(BUILD_REPOSITORY $ENV{AZURECIREPO})
 endif()
 if (DEFINED ENV{TITLEBARFORMATIDLE})
   set(TITLE_BAR_FORMAT_IDLE $ENV{TITLEBARFORMATIDLE})

src/common/bit_field.h

@@ -36,6 +36,13 @@
 #include "common/common_funcs.h"
 #include "common/swap.h"
 
+// Inlining
+#ifdef _WIN32
+#define FORCE_INLINE __forceinline
+#else
+#define FORCE_INLINE inline __attribute__((always_inline))
+#endif
+
 /*
  * Abstract bitfield class
  *
@@ -168,11 +175,11 @@ public:
     constexpr BitField(BitField&&) noexcept = default;
     constexpr BitField& operator=(BitField&&) noexcept = default;
 
-    constexpr FORCE_INLINE operator T() const {
+    constexpr operator T() const {
         return Value();
     }
 
-    constexpr FORCE_INLINE void Assign(const T& value) {
+    constexpr void Assign(const T& value) {
         storage = (static_cast<StorageType>(storage) & ~mask) | FormatValue(value);
     }
 

src/common/common_funcs.h

@@ -1,10 +1,11 @@
-// Copyright 2013 Dolphin Emulator Project / 2014 Citra Emulator Project
+// Copyright 2019 yuzu emulator team
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
 #pragma once
 
 #include <algorithm>
+#include <array>
 #include <string>
 
 #if !defined(ARCHITECTURE_x86_64)
@@ -16,18 +17,15 @@
 #define CONCAT2(x, y) DO_CONCAT2(x, y)
 #define DO_CONCAT2(x, y) x##y
 
-// helper macro to properly align structure members.
-// Calling INSERT_PADDING_BYTES will add a new member variable with a name like "pad121",
-// depending on the current source line to make sure variable names are unique.
-#define INSERT_PADDING_BYTES(num_bytes) u8 CONCAT2(pad, __LINE__)[(num_bytes)]
-#define INSERT_PADDING_WORDS(num_words) u32 CONCAT2(pad, __LINE__)[(num_words)]
+/// Helper macros to insert unused bytes or words to properly align structs. These values will be
+/// zero-initialized.
+#define INSERT_PADDING_BYTES(num_bytes) std::array<u8, num_bytes> CONCAT2(pad, __LINE__){};
+#define INSERT_PADDING_WORDS(num_words) std::array<u32, num_words> CONCAT2(pad, __LINE__){};
 
-// Inlining
-#ifdef _WIN32
-#define FORCE_INLINE __forceinline
-#else
-#define FORCE_INLINE inline __attribute__((always_inline))
-#endif
+/// These are similar to the INSERT_PADDING_* macros, but are needed for padding unions. This is
+/// because unions can only be initialized by one member.
+#define INSERT_UNION_PADDING_BYTES(num_bytes) std::array<u8, num_bytes> CONCAT2(pad, __LINE__);
+#define INSERT_UNION_PADDING_WORDS(num_words) std::array<u32, num_words> CONCAT2(pad, __LINE__);
 
 #ifndef _MSC_VER
 

src/common/multi_level_queue.h

@@ -304,6 +304,13 @@ public:
         return levels[priority == Depth ? 63 : priority].back();
     }
 
+    void clear() {
+        used_priorities = 0;
+        for (std::size_t i = 0; i < Depth; i++) {
+            levels[i].clear();
+        }
+    }
+
 private:
     using const_list_iterator = typename std::list<T>::const_iterator;
 

src/core/CMakeLists.txt

@@ -522,6 +522,23 @@ add_library(core STATIC
     tools/freezer.h
 )
 
+if (MSVC)
+    target_compile_options(core PRIVATE
+        # 'expression' : signed/unsigned mismatch
+        /we4018
+        # 'argument' : conversion from 'type1' to 'type2', possible loss of data (floating-point)
+        /we4244
+        # 'conversion' : conversion from 'type1' to 'type2', signed/unsigned mismatch
+        /we4245
+        # 'operator': conversion from 'type1:field_bits' to 'type2:field_bits', possible loss of data
+        /we4254
+        # 'var' : conversion from 'size_t' to 'type', possible loss of data
+        /we4267
+        # 'context' : truncation from 'type1' to 'type2'
+        /we4305
+    )
+endif()
+
 create_target_directory_groups(core)
 
 target_link_libraries(core PUBLIC common PRIVATE audio_core video_core)

src/core/arm/dynarmic/arm_dynarmic.cpp

@@ -67,7 +67,7 @@ public:
         ARM_Interface::ThreadContext ctx;
         parent.SaveContext(ctx);
         parent.inner_unicorn.LoadContext(ctx);
-        parent.inner_unicorn.ExecuteInstructions(static_cast<int>(num_instructions));
+        parent.inner_unicorn.ExecuteInstructions(num_instructions);
         parent.inner_unicorn.SaveContext(ctx);
         parent.LoadContext(ctx);
         num_interpreted_instructions += num_instructions;

src/core/arm/unicorn/arm_unicorn.cpp

@@ -67,10 +67,11 @@ ARM_Unicorn::ARM_Unicorn(System& system) : system{system} {
     CHECKED(uc_reg_write(uc, UC_ARM64_REG_CPACR_EL1, &fpv));
 
     uc_hook hook{};
-    CHECKED(uc_hook_add(uc, &hook, UC_HOOK_INTR, (void*)InterruptHook, this, 0, -1));
-    CHECKED(uc_hook_add(uc, &hook, UC_HOOK_MEM_INVALID, (void*)UnmappedMemoryHook, &system, 0, -1));
+    CHECKED(uc_hook_add(uc, &hook, UC_HOOK_INTR, (void*)InterruptHook, this, 0, UINT64_MAX));
+    CHECKED(uc_hook_add(uc, &hook, UC_HOOK_MEM_INVALID, (void*)UnmappedMemoryHook, &system, 0,
+                        UINT64_MAX));
     if (GDBStub::IsServerEnabled()) {
-        CHECKED(uc_hook_add(uc, &hook, UC_HOOK_CODE, (void*)CodeHook, this, 0, -1));
+        CHECKED(uc_hook_add(uc, &hook, UC_HOOK_CODE, (void*)CodeHook, this, 0, UINT64_MAX));
         last_bkpt_hit = false;
     }
 }
@@ -154,9 +155,10 @@ void ARM_Unicorn::SetTPIDR_EL0(u64 value) {
 
 void ARM_Unicorn::Run() {
     if (GDBStub::IsServerEnabled()) {
-        ExecuteInstructions(std::max(4000000, 0));
+        ExecuteInstructions(std::max(4000000U, 0U));
     } else {
-        ExecuteInstructions(std::max(system.CoreTiming().GetDowncount(), s64{0}));
+        ExecuteInstructions(
+            std::max(std::size_t(system.CoreTiming().GetDowncount()), std::size_t{0}));
     }
 }
 
@@ -166,7 +168,7 @@ void ARM_Unicorn::Step() {
 
 MICROPROFILE_DEFINE(ARM_Jit_Unicorn, "ARM JIT", "Unicorn", MP_RGB(255, 64, 64));
 
-void ARM_Unicorn::ExecuteInstructions(int num_instructions) {
+void ARM_Unicorn::ExecuteInstructions(std::size_t num_instructions) {
     MICROPROFILE_SCOPE(ARM_Jit_Unicorn);
     CHECKED(uc_emu_start(uc, GetPC(), 1ULL << 63, 0, num_instructions));
     system.CoreTiming().AddTicks(num_instructions);

src/core/arm/unicorn/arm_unicorn.h

@@ -34,7 +34,7 @@ public:
     void LoadContext(const ThreadContext& ctx) override;
     void PrepareReschedule() override;
     void ClearExclusiveState() override;
-    void ExecuteInstructions(int num_instructions);
+    void ExecuteInstructions(std::size_t num_instructions);
     void Run() override;
     void Step() override;
     void ClearInstructionCache() override;

src/core/core.cpp

@@ -409,6 +409,12 @@ void System::PrepareReschedule() {
     CurrentCpuCore().PrepareReschedule();
 }
 
+void System::PrepareReschedule(const u32 core_index) {
+    if (core_index < GlobalScheduler().CpuCoresCount()) {
+        CpuCore(core_index).PrepareReschedule();
+    }
+}
+
 PerfStatsResults System::GetAndResetPerfStats() {
     return impl->GetAndResetPerfStats();
 }
@@ -449,6 +455,16 @@ const Kernel::Scheduler& System::Scheduler(std::size_t core_index) const {
     return CpuCore(core_index).Scheduler();
 }
 
+/// Gets the global scheduler
+Kernel::GlobalScheduler& System::GlobalScheduler() {
+    return impl->kernel.GlobalScheduler();
+}
+
+/// Gets the global scheduler
+const Kernel::GlobalScheduler& System::GlobalScheduler() const {
+    return impl->kernel.GlobalScheduler();
+}
+
 Kernel::Process* System::CurrentProcess() {
     return impl->kernel.CurrentProcess();
 }

src/core/core.h

@@ -24,6 +24,7 @@ class VfsFilesystem;
 } // namespace FileSys
 
 namespace Kernel {
+class GlobalScheduler;
 class KernelCore;
 class Process;
 class Scheduler;
@@ -184,6 +185,9 @@ public:
     /// Prepare the core emulation for a reschedule
     void PrepareReschedule();
 
+    /// Prepare the core emulation for a reschedule
+    void PrepareReschedule(u32 core_index);
+
     /// Gets and resets core performance statistics
     PerfStatsResults GetAndResetPerfStats();
 
@@ -238,6 +242,12 @@ public:
     /// Gets the scheduler for the CPU core with the specified index
     const Kernel::Scheduler& Scheduler(std::size_t core_index) const;
 
+    /// Gets the global scheduler
+    Kernel::GlobalScheduler& GlobalScheduler();
+
+    /// Gets the global scheduler
+    const Kernel::GlobalScheduler& GlobalScheduler() const;
+
     /// Provides a pointer to the current process
     Kernel::Process* CurrentProcess();
 

src/core/core_cpu.cpp

@@ -52,7 +52,8 @@ bool CpuBarrier::Rendezvous() {
 
 Cpu::Cpu(System& system, ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_barrier,
          std::size_t core_index)
-    : cpu_barrier{cpu_barrier}, core_timing{system.CoreTiming()}, core_index{core_index} {
+    : cpu_barrier{cpu_barrier}, global_scheduler{system.GlobalScheduler()},
+      core_timing{system.CoreTiming()}, core_index{core_index} {
 #ifdef ARCHITECTURE_x86_64
     arm_interface = std::make_unique<ARM_Dynarmic>(system, exclusive_monitor, core_index);
 #else
@@ -60,7 +61,7 @@ Cpu::Cpu(System& system, ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_ba
     LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
 #endif
 
-    scheduler = std::make_unique<Kernel::Scheduler>(system, *arm_interface);
+    scheduler = std::make_unique<Kernel::Scheduler>(system, *arm_interface, core_index);
 }
 
 Cpu::~Cpu() = default;
@@ -81,21 +82,21 @@ void Cpu::RunLoop(bool tight_loop) {
         return;
     }
 
+    Reschedule();
+
     // If we don't have a currently active thread then don't execute instructions,
     // instead advance to the next event and try to yield to the next thread
     if (Kernel::GetCurrentThread() == nullptr) {
         LOG_TRACE(Core, "Core-{} idling", core_index);
         core_timing.Idle();
-        core_timing.Advance();
-        PrepareReschedule();
     } else {
         if (tight_loop) {
             arm_interface->Run();
         } else {
             arm_interface->Step();
         }
-        core_timing.Advance();
     }
+    core_timing.Advance();
 
     Reschedule();
 }
@@ -106,18 +107,18 @@ void Cpu::SingleStep() {
 
 void Cpu::PrepareReschedule() {
     arm_interface->PrepareReschedule();
-    reschedule_pending = true;
 }
 
 void Cpu::Reschedule() {
-    if (!reschedule_pending) {
-        return;
-    }
-
-    reschedule_pending = false;
     // Lock the global kernel mutex when we manipulate the HLE state
-    std::lock_guard lock{HLE::g_hle_lock};
-    scheduler->Reschedule();
+    std::lock_guard lock(HLE::g_hle_lock);
+
+    global_scheduler.SelectThread(core_index);
+    scheduler->TryDoContextSwitch();
+}
+
+void Cpu::Shutdown() {
+    scheduler->Shutdown();
 }
 
 } // namespace Core

src/core/core_cpu.h

@@ -12,8 +12,9 @@
 #include "common/common_types.h"
 
 namespace Kernel {
+class GlobalScheduler;
 class Scheduler;
-}
+} // namespace Kernel
 
 namespace Core {
 class System;
@@ -83,6 +84,8 @@ public:
         return core_index;
     }
 
+    void Shutdown();
+
     static std::unique_ptr<ExclusiveMonitor> MakeExclusiveMonitor(std::size_t num_cores);
 
 private:
@@ -90,6 +93,7 @@ private:
 
     std::unique_ptr<ARM_Interface> arm_interface;
     CpuBarrier& cpu_barrier;
+    Kernel::GlobalScheduler& global_scheduler;
     std::unique_ptr<Kernel::Scheduler> scheduler;
     Timing::CoreTiming& core_timing;
 

src/core/cpu_core_manager.cpp

@@ -58,6 +58,7 @@ void CpuCoreManager::Shutdown() {
 
     thread_to_cpu.clear();
     for (auto& cpu_core : cores) {
+        cpu_core->Shutdown();
         cpu_core.reset();
     }
 

src/core/crypto/key_manager.cpp

@@ -22,6 +22,7 @@
 #include "common/file_util.h"
 #include "common/hex_util.h"
 #include "common/logging/log.h"
+#include "common/string_util.h"
 #include "core/core.h"
 #include "core/crypto/aes_util.h"
 #include "core/crypto/key_manager.h"
@@ -378,8 +379,9 @@ std::vector<Ticket> GetTicketblob(const FileUtil::IOFile& ticket_save) {
 template <size_t size>
 static std::array<u8, size> operator^(const std::array<u8, size>& lhs,
                                       const std::array<u8, size>& rhs) {
-    std::array<u8, size> out{};
-    std::transform(lhs.begin(), lhs.end(), rhs.begin(), out.begin(), std::bit_xor<>());
+    std::array<u8, size> out;
+    std::transform(lhs.begin(), lhs.end(), rhs.begin(), out.begin(),
+                   [](u8 lhs, u8 rhs) { return u8(lhs ^ rhs); });
     return out;
 }
 
@@ -538,7 +540,7 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) {
             Key128 key = Common::HexStringToArray<16>(out[1]);
             s128_keys[{S128KeyType::Titlekey, rights_id[1], rights_id[0]}] = key;
         } else {
-            std::transform(out[0].begin(), out[0].end(), out[0].begin(), ::tolower);
+            out[0] = Common::ToLower(out[0]);
             if (s128_file_id.find(out[0]) != s128_file_id.end()) {
                 const auto index = s128_file_id.at(out[0]);
                 Key128 key = Common::HexStringToArray<16>(out[1]);
@@ -944,12 +946,10 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) {
         return;
     }
 
-    Key128 rsa_oaep_kek{};
-    std::transform(seed3.begin(), seed3.end(), mask0.begin(), rsa_oaep_kek.begin(),
-                   std::bit_xor<>());
-
-    if (rsa_oaep_kek == Key128{})
+    const Key128 rsa_oaep_kek = seed3 ^ mask0;
+    if (rsa_oaep_kek == Key128{}) {
         return;
+    }
 
     SetKey(S128KeyType::Source, rsa_oaep_kek,
            static_cast<u64>(SourceKeyType::RSAOaepKekGeneration));

src/core/crypto/partition_data_manager.cpp

@@ -204,11 +204,12 @@ static std::array<Key128, 0x20> FindEncryptedMasterKeyFromHex(const std::vector<
 
 FileSys::VirtualFile FindFileInDirWithNames(const FileSys::VirtualDir& dir,
                                             const std::string& name) {
-    auto upper = name;
-    std::transform(upper.begin(), upper.end(), upper.begin(), [](u8 c) { return std::toupper(c); });
+    const auto upper = Common::ToUpper(name);
+
     for (const auto& fname : {name, name + ".bin", upper, upper + ".BIN"}) {
-        if (dir->GetFile(fname) != nullptr)
+        if (dir->GetFile(fname) != nullptr) {
             return dir->GetFile(fname);
+        }
     }
 
     return nullptr;

src/core/file_sys/content_archive.cpp

@@ -32,11 +32,28 @@ enum class NCASectionFilesystemType : u8 {
     ROMFS = 0x3,
 };
 
+struct IVFCLevel {
+    u64_le offset;
+    u64_le size;
+    u32_le block_size;
+    u32_le reserved;
+};
+static_assert(sizeof(IVFCLevel) == 0x18, "IVFCLevel has incorrect size.");
+
+struct IVFCHeader {
+    u32_le magic;
+    u32_le magic_number;
+    INSERT_UNION_PADDING_BYTES(8);
+    std::array<IVFCLevel, 6> levels;
+    INSERT_UNION_PADDING_BYTES(64);
+};
+static_assert(sizeof(IVFCHeader) == 0xE0, "IVFCHeader has incorrect size.");
+
 struct NCASectionHeaderBlock {
-    INSERT_PADDING_BYTES(3);
+    INSERT_UNION_PADDING_BYTES(3);
     NCASectionFilesystemType filesystem_type;
     NCASectionCryptoType crypto_type;
-    INSERT_PADDING_BYTES(3);
+    INSERT_UNION_PADDING_BYTES(3);
 };
 static_assert(sizeof(NCASectionHeaderBlock) == 0x8, "NCASectionHeaderBlock has incorrect size.");
 
@@ -44,7 +61,7 @@ struct NCASectionRaw {
     NCASectionHeaderBlock header;
     std::array<u8, 0x138> block_data;
     std::array<u8, 0x8> section_ctr;
-    INSERT_PADDING_BYTES(0xB8);
+    INSERT_UNION_PADDING_BYTES(0xB8);
 };
 static_assert(sizeof(NCASectionRaw) == 0x200, "NCASectionRaw has incorrect size.");
 
@@ -52,19 +69,19 @@ struct PFS0Superblock {
     NCASectionHeaderBlock header_block;
     std::array<u8, 0x20> hash;
     u32_le size;
-    INSERT_PADDING_BYTES(4);
+    INSERT_UNION_PADDING_BYTES(4);
     u64_le hash_table_offset;
     u64_le hash_table_size;
     u64_le pfs0_header_offset;
     u64_le pfs0_size;
-    INSERT_PADDING_BYTES(0x1B0);
+    INSERT_UNION_PADDING_BYTES(0x1B0);
 };
 static_assert(sizeof(PFS0Superblock) == 0x200, "PFS0Superblock has incorrect size.");
 
 struct RomFSSuperblock {
     NCASectionHeaderBlock header_block;
     IVFCHeader ivfc;
-    INSERT_PADDING_BYTES(0x118);
+    INSERT_UNION_PADDING_BYTES(0x118);
 };
 static_assert(sizeof(RomFSSuperblock) == 0x200, "RomFSSuperblock has incorrect size.");
 
@@ -72,24 +89,24 @@ struct BKTRHeader {
     u64_le offset;
     u64_le size;
     u32_le magic;
-    INSERT_PADDING_BYTES(0x4);
+    INSERT_UNION_PADDING_BYTES(0x4);
     u32_le number_entries;
-    INSERT_PADDING_BYTES(0x4);
+    INSERT_UNION_PADDING_BYTES(0x4);
 };
 static_assert(sizeof(BKTRHeader) == 0x20, "BKTRHeader has incorrect size.");
 
 struct BKTRSuperblock {
     NCASectionHeaderBlock header_block;
     IVFCHeader ivfc;
-    INSERT_PADDING_BYTES(0x18);
+    INSERT_UNION_PADDING_BYTES(0x18);
     BKTRHeader relocation;
     BKTRHeader subsection;
-    INSERT_PADDING_BYTES(0xC0);
+    INSERT_UNION_PADDING_BYTES(0xC0);
 };
 static_assert(sizeof(BKTRSuperblock) == 0x200, "BKTRSuperblock has incorrect size.");
 
 union NCASectionHeader {
-    NCASectionRaw raw;
+    NCASectionRaw raw{};
     PFS0Superblock pfs0;
     RomFSSuperblock romfs;
     BKTRSuperblock bktr;

src/core/file_sys/kernel_executable.cpp

@@ -147,7 +147,7 @@ std::vector<u32> KIP::GetKernelCapabilities() const {
 }
 
 s32 KIP::GetMainThreadPriority() const {
-    return header.main_thread_priority;
+    return static_cast<s32>(header.main_thread_priority);
 }
 
 u32 KIP::GetMainThreadStackSize() const {

src/core/file_sys/program_metadata.cpp

@@ -52,14 +52,14 @@ Loader::ResultStatus ProgramMetadata::Load(VirtualFile file) {
 }
 
 void ProgramMetadata::LoadManual(bool is_64_bit, ProgramAddressSpaceType address_space,
-                                 u8 main_thread_prio, u8 main_thread_core,
+                                 s32 main_thread_prio, u32 main_thread_core,
                                  u32 main_thread_stack_size, u64 title_id,
                                  u64 filesystem_permissions,
                                  KernelCapabilityDescriptors capabilities) {
     npdm_header.has_64_bit_instructions.Assign(is_64_bit);
     npdm_header.address_space_type.Assign(address_space);
-    npdm_header.main_thread_priority = main_thread_prio;
-    npdm_header.main_thread_cpu = main_thread_core;
+    npdm_header.main_thread_priority = static_cast<u8>(main_thread_prio);
+    npdm_header.main_thread_cpu = static_cast<u8>(main_thread_core);
     npdm_header.main_stack_size = main_thread_stack_size;
     aci_header.title_id = title_id;
     aci_file_access.permissions = filesystem_permissions;

src/core/file_sys/program_metadata.h

@@ -47,8 +47,8 @@ public:
     Loader::ResultStatus Load(VirtualFile file);
 
     // Load from parameters instead of NPDM file, used for KIP
-    void LoadManual(bool is_64_bit, ProgramAddressSpaceType address_space, u8 main_thread_prio,
-                    u8 main_thread_core, u32 main_thread_stack_size, u64 title_id,
+    void LoadManual(bool is_64_bit, ProgramAddressSpaceType address_space, s32 main_thread_prio,
+                    u32 main_thread_core, u32 main_thread_stack_size, u64 title_id,
                     u64 filesystem_permissions, KernelCapabilityDescriptors capabilities);
 
     bool Is64BitProgram() const;

src/core/file_sys/romfs.h

@@ -13,25 +13,6 @@
 
 namespace FileSys {
 
-struct RomFSHeader;
-
-struct IVFCLevel {
-    u64_le offset;
-    u64_le size;
-    u32_le block_size;
-    u32_le reserved;
-};
-static_assert(sizeof(IVFCLevel) == 0x18, "IVFCLevel has incorrect size.");
-
-struct IVFCHeader {
-    u32_le magic;
-    u32_le magic_number;
-    INSERT_PADDING_BYTES(8);
-    std::array<IVFCLevel, 6> levels;
-    INSERT_PADDING_BYTES(64);
-};
-static_assert(sizeof(IVFCHeader) == 0xE0, "IVFCHeader has incorrect size.");
-
 enum class RomFSExtractionType {
     Full,          // Includes data directory
     Truncated,     // Traverses into data directory

src/core/file_sys/romfs_factory.cpp

@@ -71,12 +71,12 @@ ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id, StorageId storage,
 
     if (res == nullptr) {
         // TODO(DarkLordZach): Find the right error code to use here
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
     const auto romfs = res->GetRomFS();
     if (romfs == nullptr) {
         // TODO(DarkLordZach): Find the right error code to use here
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
     return MakeResult<VirtualFile>(romfs);
 }

src/core/file_sys/savedata_factory.cpp

@@ -16,6 +16,7 @@ namespace FileSys {
 constexpr char SAVE_DATA_SIZE_FILENAME[] = ".yuzu_save_size";
 
 namespace {
+
 void PrintSaveDataDescriptorWarnings(SaveDataDescriptor meta) {
     if (meta.type == SaveDataType::SystemSaveData || meta.type == SaveDataType::SaveData) {
         if (meta.zero_1 != 0) {
@@ -52,6 +53,13 @@ void PrintSaveDataDescriptorWarnings(SaveDataDescriptor meta) {
                     meta.user_id[1], meta.user_id[0]);
     }
 }
+
+bool ShouldSaveDataBeAutomaticallyCreated(SaveDataSpaceId space, const SaveDataDescriptor& desc) {
+    return desc.type == SaveDataType::CacheStorage || desc.type == SaveDataType::TemporaryStorage ||
+           (space == SaveDataSpaceId::NandUser && ///< Normal Save Data -- Current Title & User
+            desc.type == SaveDataType::SaveData && desc.title_id == 0 && desc.save_id == 0);
+}
+
 } // Anonymous namespace
 
 std::string SaveDataDescriptor::DebugInfo() const {
@@ -82,7 +90,7 @@ ResultVal<VirtualDir> SaveDataFactory::Create(SaveDataSpaceId space,
     // Return an error if the save data doesn't actually exist.
     if (out == nullptr) {
         // TODO(DarkLordZach): Find out correct error code.
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
 
     return MakeResult<VirtualDir>(std::move(out));
@@ -96,10 +104,14 @@ ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space,
 
     auto out = dir->GetDirectoryRelative(save_directory);
 
+    if (out == nullptr && ShouldSaveDataBeAutomaticallyCreated(space, meta)) {
+        return Create(space, meta);
+    }
+
     // Return an error if the save data doesn't actually exist.
     if (out == nullptr) {
         // TODO(Subv): Find out correct error code.
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
 
     return MakeResult<VirtualDir>(std::move(out));

src/core/file_sys/vfs_libzip.cpp

@@ -27,7 +27,7 @@ VirtualDir ExtractZIP(VirtualFile file) {
 
     std::shared_ptr<VectorVfsDirectory> out = std::make_shared<VectorVfsDirectory>();
 
-    const auto num_entries = zip_get_num_entries(zip.get(), 0);
+    const auto num_entries = static_cast<std::size_t>(zip_get_num_entries(zip.get(), 0));
 
     zip_stat_t stat{};
     zip_stat_init(&stat);

src/core/file_sys/xts_archive.cpp

@@ -7,12 +7,13 @@
 #include <cstring>
 #include <regex>
 #include <string>
+
 #include <mbedtls/md.h>
 #include <mbedtls/sha256.h>
-#include "common/assert.h"
+
 #include "common/file_util.h"
 #include "common/hex_util.h"
-#include "common/logging/log.h"
+#include "common/string_util.h"
 #include "core/crypto/aes_util.h"
 #include "core/crypto/xts_encryption_layer.h"
 #include "core/file_sys/partition_filesystem.h"
@@ -53,11 +54,8 @@ NAX::NAX(VirtualFile file_) : header(std::make_unique<NAXHeader>()), file(std::m
         return;
     }
 
-    std::string two_dir = match[1];
-    std::string nca_id = match[2];
-    std::transform(two_dir.begin(), two_dir.end(), two_dir.begin(), ::toupper);
-    std::transform(nca_id.begin(), nca_id.end(), nca_id.begin(), ::tolower);
-
+    const std::string two_dir = Common::ToUpper(match[1]);
+    const std::string nca_id = Common::ToLower(match[2]);
     status = Parse(fmt::format("/registered/{}/{}.nca", two_dir, nca_id));
 }
 

src/core/gdbstub/gdbstub.cpp

@@ -202,13 +202,11 @@ void RegisterModule(std::string name, VAddr beg, VAddr end, bool add_elf_ext) {
 }
 
 static Kernel::Thread* FindThreadById(s64 id) {
-    for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
-        const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
-        for (auto& thread : threads) {
-            if (thread->GetThreadID() == static_cast<u64>(id)) {
-                current_core = core;
-                return thread.get();
-            }
+    const auto& threads = Core::System::GetInstance().GlobalScheduler().GetThreadList();
+    for (auto& thread : threads) {
+        if (thread->GetThreadID() == static_cast<u64>(id)) {
+            current_core = thread->GetProcessorID();
+            return thread.get();
         }
     }
     return nullptr;
@@ -470,7 +468,8 @@ static u8 ReadByte() {
 
 /// Calculate the checksum of the current command buffer.
 static u8 CalculateChecksum(const u8* buffer, std::size_t length) {
-    return static_cast<u8>(std::accumulate(buffer, buffer + length, 0, std::plus<u8>()));
+    return static_cast<u8>(std::accumulate(buffer, buffer + length, u8{0},
+                                           [](u8 lhs, u8 rhs) { return u8(lhs + rhs); }));
 }
 
 /**
@@ -647,11 +646,9 @@ static void HandleQuery() {
         SendReply(buffer.c_str());
     } else if (strncmp(query, "fThreadInfo", strlen("fThreadInfo")) == 0) {
         std::string val = "m";
-        for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
-            const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
-            for (const auto& thread : threads) {
-                val += fmt::format("{:x},", thread->GetThreadID());
-            }
+        const auto& threads = Core::System::GetInstance().GlobalScheduler().GetThreadList();
+        for (const auto& thread : threads) {
+            val += fmt::format("{:x},", thread->GetThreadID());
         }
         val.pop_back();
         SendReply(val.c_str());
@@ -661,13 +658,11 @@ static void HandleQuery() {
         std::string buffer;
         buffer += "l<?xml version=\"1.0\"?>";
         buffer += "<threads>";
-        for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
-            const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
-            for (const auto& thread : threads) {
-                buffer +=
-                    fmt::format(R"*(<thread id="{:x}" core="{:d}" name="Thread {:x}"></thread>)*",
-                                thread->GetThreadID(), core, thread->GetThreadID());
-            }
+        const auto& threads = Core::System::GetInstance().GlobalScheduler().GetThreadList();
+        for (const auto& thread : threads) {
+            buffer +=
+                fmt::format(R"*(<thread id="{:x}" core="{:d}" name="Thread {:x}"></thread>)*",
+                            thread->GetThreadID(), thread->GetProcessorID(), thread->GetThreadID());
         }
         buffer += "</threads>";
         SendReply(buffer.c_str());

src/core/hle/ipc.h

@@ -160,7 +160,7 @@ struct DomainMessageHeader {
         // Used when responding to an IPC request, Server -> Client.
         struct {
             u32_le num_objects;
-            INSERT_PADDING_WORDS(3);
+            INSERT_UNION_PADDING_WORDS(3);
         };
 
         // Used when performing an IPC request, Client -> Server.
@@ -171,8 +171,10 @@ struct DomainMessageHeader {
                 BitField<16, 16, u32> size;
             };
             u32_le object_id;
-            INSERT_PADDING_WORDS(2);
+            INSERT_UNION_PADDING_WORDS(2);
         };
+
+        std::array<u32, 4> raw{};
     };
 };
 static_assert(sizeof(DomainMessageHeader) == 16, "DomainMessageHeader size is incorrect");

src/core/hle/kernel/address_arbiter.cpp

@@ -22,6 +22,7 @@ namespace Kernel {
 namespace {
 // Wake up num_to_wake (or all) threads in a vector.
 void WakeThreads(const std::vector<SharedPtr<Thread>>& waiting_threads, s32 num_to_wake) {
+    auto& system = Core::System::GetInstance();
     // Only process up to 'target' threads, unless 'target' is <= 0, in which case process
     // them all.
     std::size_t last = waiting_threads.size();
@@ -35,6 +36,7 @@ void WakeThreads(const std::vector<SharedPtr<Thread>>& waiting_threads, s32 num_
         waiting_threads[i]->SetWaitSynchronizationResult(RESULT_SUCCESS);
         waiting_threads[i]->SetArbiterWaitAddress(0);
         waiting_threads[i]->ResumeFromWait();
+        system.PrepareReschedule(waiting_threads[i]->GetProcessorID());
     }
 }
 } // Anonymous namespace
@@ -89,12 +91,20 @@ ResultCode AddressArbiter::ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr a
 
     // Determine the modified value depending on the waiting count.
     s32 updated_value;
-    if (waiting_threads.empty()) {
-        updated_value = value + 1;
-    } else if (num_to_wake <= 0 || waiting_threads.size() <= static_cast<u32>(num_to_wake)) {
-        updated_value = value - 1;
+    if (num_to_wake <= 0) {
+        if (waiting_threads.empty()) {
+            updated_value = value + 1;
+        } else {
+            updated_value = value - 1;
+        }
     } else {
-        updated_value = value;
+        if (waiting_threads.empty()) {
+            updated_value = value + 1;
+        } else if (waiting_threads.size() <= static_cast<u32>(num_to_wake)) {
+            updated_value = value - 1;
+        } else {
+            updated_value = value;
+        }
     }
 
     if (static_cast<s32>(Memory::Read32(address)) != value) {
@@ -169,30 +179,22 @@ ResultCode AddressArbiter::WaitForAddressImpl(VAddr address, s64 timeout) {
 
     current_thread->WakeAfterDelay(timeout);
 
-    system.CpuCore(current_thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(current_thread->GetProcessorID());
     return RESULT_TIMEOUT;
 }
 
 std::vector<SharedPtr<Thread>> AddressArbiter::GetThreadsWaitingOnAddress(VAddr address) const {
-    const auto RetrieveWaitingThreads = [this](std::size_t core_index,
-                                               std::vector<SharedPtr<Thread>>& waiting_threads,
-                                               VAddr arb_addr) {
-        const auto& scheduler = system.Scheduler(core_index);
-        const auto& thread_list = scheduler.GetThreadList();
-
-        for (const auto& thread : thread_list) {
-            if (thread->GetArbiterWaitAddress() == arb_addr) {
-                waiting_threads.push_back(thread);
-            }
-        }
-    };
 
     // Retrieve all threads that are waiting for this address.
     std::vector<SharedPtr<Thread>> threads;
-    RetrieveWaitingThreads(0, threads, address);
-    RetrieveWaitingThreads(1, threads, address);
-    RetrieveWaitingThreads(2, threads, address);
-    RetrieveWaitingThreads(3, threads, address);
+    const auto& scheduler = system.GlobalScheduler();
+    const auto& thread_list = scheduler.GetThreadList();
+
+    for (const auto& thread : thread_list) {
+        if (thread->GetArbiterWaitAddress() == address) {
+            threads.push_back(thread);
+        }
+    }
 
     // Sort them by priority, such that the highest priority ones come first.
     std::sort(threads.begin(), threads.end(),

src/core/hle/kernel/hle_ipc.cpp

@@ -58,8 +58,7 @@ SharedPtr<WritableEvent> HLERequestContext::SleepClientThread(
     auto& kernel = Core::System::GetInstance().Kernel();
     if (!writable_event) {
         // Create event if not provided
-        const auto pair = WritableEvent::CreateEventPair(kernel, ResetType::Automatic,
-                                                         "HLE Pause Event: " + reason);
+        const auto pair = WritableEvent::CreateEventPair(kernel, "HLE Pause Event: " + reason);
         writable_event = pair.writable;
     }
 

src/core/hle/kernel/kernel.cpp

@@ -12,12 +12,15 @@
 
 #include "core/core.h"
 #include "core/core_timing.h"
+#include "core/core_timing_util.h"
 #include "core/hle/kernel/address_arbiter.h"
 #include "core/hle/kernel/client_port.h"
+#include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/resource_limit.h"
+#include "core/hle/kernel/scheduler.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/lock.h"
 #include "core/hle/result.h"
@@ -58,12 +61,8 @@ static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] s64 cycles_
         if (thread->HasWakeupCallback()) {
             resume = thread->InvokeWakeupCallback(ThreadWakeupReason::Timeout, thread, nullptr, 0);
         }
-    }
-
-    if (thread->GetMutexWaitAddress() != 0 || thread->GetCondVarWaitAddress() != 0 ||
-        thread->GetWaitHandle() != 0) {
-        ASSERT(thread->GetStatus() == ThreadStatus::WaitMutex ||
-               thread->GetStatus() == ThreadStatus::WaitCondVar);
+    } else if (thread->GetStatus() == ThreadStatus::WaitMutex ||
+               thread->GetStatus() == ThreadStatus::WaitCondVar) {
         thread->SetMutexWaitAddress(0);
         thread->SetCondVarWaitAddress(0);
         thread->SetWaitHandle(0);
@@ -83,18 +82,23 @@ static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] s64 cycles_
     }
 
     if (resume) {
+        if (thread->GetStatus() == ThreadStatus::WaitCondVar ||
+            thread->GetStatus() == ThreadStatus::WaitArb) {
+            thread->SetWaitSynchronizationResult(RESULT_TIMEOUT);
+        }
         thread->ResumeFromWait();
     }
 }
 
 struct KernelCore::Impl {
-    explicit Impl(Core::System& system) : system{system} {}
+    explicit Impl(Core::System& system) : system{system}, global_scheduler{system} {}
 
     void Initialize(KernelCore& kernel) {
         Shutdown();
 
         InitializeSystemResourceLimit(kernel);
         InitializeThreads();
+        InitializePreemption();
     }
 
     void Shutdown() {
@@ -110,6 +114,9 @@ struct KernelCore::Impl {
 
         thread_wakeup_callback_handle_table.Clear();
         thread_wakeup_event_type = nullptr;
+        preemption_event = nullptr;
+
+        global_scheduler.Shutdown();
 
         named_ports.clear();
     }
@@ -132,6 +139,18 @@ struct KernelCore::Impl {
             system.CoreTiming().RegisterEvent("ThreadWakeupCallback", ThreadWakeupCallback);
     }
 
+    void InitializePreemption() {
+        preemption_event = system.CoreTiming().RegisterEvent(
+            "PreemptionCallback", [this](u64 userdata, s64 cycles_late) {
+                global_scheduler.PreemptThreads();
+                s64 time_interval = Core::Timing::msToCycles(std::chrono::milliseconds(10));
+                system.CoreTiming().ScheduleEvent(time_interval, preemption_event);
+            });
+
+        s64 time_interval = Core::Timing::msToCycles(std::chrono::milliseconds(10));
+        system.CoreTiming().ScheduleEvent(time_interval, preemption_event);
+    }
+
     std::atomic<u32> next_object_id{0};
     std::atomic<u64> next_kernel_process_id{Process::InitialKIPIDMin};
     std::atomic<u64> next_user_process_id{Process::ProcessIDMin};
@@ -140,10 +159,12 @@ struct KernelCore::Impl {
     // Lists all processes that exist in the current session.
     std::vector<SharedPtr<Process>> process_list;
     Process* current_process = nullptr;
+    Kernel::GlobalScheduler global_scheduler;
 
     SharedPtr<ResourceLimit> system_resource_limit;
 
     Core::Timing::EventType* thread_wakeup_event_type = nullptr;
+    Core::Timing::EventType* preemption_event = nullptr;
     // TODO(yuriks): This can be removed if Thread objects are explicitly pooled in the future,
     // allowing us to simply use a pool index or similar.
     Kernel::HandleTable thread_wakeup_callback_handle_table;
@@ -203,6 +224,14 @@ const std::vector<SharedPtr<Process>>& KernelCore::GetProcessList() const {
     return impl->process_list;
 }
 
+Kernel::GlobalScheduler& KernelCore::GlobalScheduler() {
+    return impl->global_scheduler;
+}
+
+const Kernel::GlobalScheduler& KernelCore::GlobalScheduler() const {
+    return impl->global_scheduler;
+}
+
 void KernelCore::AddNamedPort(std::string name, SharedPtr<ClientPort> port) {
     impl->named_ports.emplace(std::move(name), std::move(port));
 }

src/core/hle/kernel/kernel.h

@@ -21,6 +21,7 @@ namespace Kernel {
 
 class AddressArbiter;
 class ClientPort;
+class GlobalScheduler;
 class HandleTable;
 class Process;
 class ResourceLimit;
@@ -75,6 +76,12 @@ public:
     /// Retrieves the list of processes.
     const std::vector<SharedPtr<Process>>& GetProcessList() const;
 
+    /// Gets the sole instance of the global scheduler
+    Kernel::GlobalScheduler& GlobalScheduler();
+
+    /// Gets the sole instance of the global scheduler
+    const Kernel::GlobalScheduler& GlobalScheduler() const;
+
     /// Adds a port to the named port table
     void AddNamedPort(std::string name, SharedPtr<ClientPort> port);
 

src/core/hle/kernel/mutex.cpp

@@ -139,6 +139,9 @@ ResultCode Mutex::Release(VAddr address) {
     thread->SetCondVarWaitAddress(0);
     thread->SetMutexWaitAddress(0);
     thread->SetWaitHandle(0);
+    thread->SetWaitSynchronizationResult(RESULT_SUCCESS);
+
+    system.PrepareReschedule();
 
     return RESULT_SUCCESS;
 }

src/core/hle/kernel/object.h

@@ -32,11 +32,6 @@ enum class HandleType : u32 {
     ServerSession,
 };
 
-enum class ResetType {
-    Automatic, ///< Reset automatically on object acquisition
-    Manual,    ///< Never reset automatically
-};
-
 class Object : NonCopyable {
 public:
     explicit Object(KernelCore& kernel);

src/core/hle/kernel/process.cpp

@@ -213,10 +213,7 @@ void Process::PrepareForTermination() {
         }
     };
 
-    stop_threads(system.Scheduler(0).GetThreadList());
-    stop_threads(system.Scheduler(1).GetThreadList());
-    stop_threads(system.Scheduler(2).GetThreadList());
-    stop_threads(system.Scheduler(3).GetThreadList());
+    stop_threads(system.GlobalScheduler().GetThreadList());
 
     FreeTLSRegion(tls_region_address);
     tls_region_address = 0;

src/core/hle/kernel/readable_event.cpp

@@ -20,15 +20,13 @@ bool ReadableEvent::ShouldWait(const Thread* thread) const {
 
 void ReadableEvent::Acquire(Thread* thread) {
     ASSERT_MSG(!ShouldWait(thread), "object unavailable!");
-
-    if (reset_type == ResetType::Automatic) {
-        signaled = false;
-    }
 }
 
 void ReadableEvent::Signal() {
-    signaled = true;
-    WakeupAllWaitingThreads();
+    if (!signaled) {
+        signaled = true;
+        WakeupAllWaitingThreads();
+    };
 }
 
 void ReadableEvent::Clear() {

src/core/hle/kernel/readable_event.h

@@ -27,10 +27,6 @@ public:
         return name;
     }
 
-    ResetType GetResetType() const {
-        return reset_type;
-    }
-
     static constexpr HandleType HANDLE_TYPE = HandleType::ReadableEvent;
     HandleType GetHandleType() const override {
         return HANDLE_TYPE;
@@ -55,8 +51,7 @@ private:
 
     void Signal();
 
-    ResetType reset_type;
-    bool signaled;
+    bool signaled{};
 
     std::string name; ///< Name of event (optional)
 };

src/core/hle/kernel/scheduler.cpp

@@ -1,8 +1,13 @@
 // Copyright 2018 yuzu emulator team
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+//
+// SelectThreads, Yield functions originally by TuxSH.
+// licensed under GPLv2 or later under exception provided by the author.
 
 #include <algorithm>
+#include <set>
+#include <unordered_set>
 #include <utility>
 
 #include "common/assert.h"
@@ -17,56 +22,374 @@
 
 namespace Kernel {
 
-std::mutex Scheduler::scheduler_mutex;
+GlobalScheduler::GlobalScheduler(Core::System& system) : system{system} {}
 
-Scheduler::Scheduler(Core::System& system, Core::ARM_Interface& cpu_core)
-    : cpu_core{cpu_core}, system{system} {}
+GlobalScheduler::~GlobalScheduler() = default;
 
-Scheduler::~Scheduler() {
-    for (auto& thread : thread_list) {
-        thread->Stop();
+void GlobalScheduler::AddThread(SharedPtr<Thread> thread) {
+    thread_list.push_back(std::move(thread));
+}
+
+void GlobalScheduler::RemoveThread(const Thread* thread) {
+    thread_list.erase(std::remove(thread_list.begin(), thread_list.end(), thread),
+                      thread_list.end());
+}
+
+void GlobalScheduler::UnloadThread(std::size_t core) {
+    Scheduler& sched = system.Scheduler(core);
+    sched.UnloadThread();
+}
+
+void GlobalScheduler::SelectThread(std::size_t core) {
+    const auto update_thread = [](Thread* thread, Scheduler& sched) {
+        if (thread != sched.selected_thread) {
+            if (thread == nullptr) {
+                ++sched.idle_selection_count;
+            }
+            sched.selected_thread = thread;
+        }
+        sched.is_context_switch_pending = sched.selected_thread != sched.current_thread;
+        std::atomic_thread_fence(std::memory_order_seq_cst);
+    };
+    Scheduler& sched = system.Scheduler(core);
+    Thread* current_thread = nullptr;
+    // Step 1: Get top thread in schedule queue.
+    current_thread = scheduled_queue[core].empty() ? nullptr : scheduled_queue[core].front();
+    if (current_thread) {
+        update_thread(current_thread, sched);
+        return;
+    }
+    // Step 2: Try selecting a suggested thread.
+    Thread* winner = nullptr;
+    std::set<s32> sug_cores;
+    for (auto thread : suggested_queue[core]) {
+        s32 this_core = thread->GetProcessorID();
+        Thread* thread_on_core = nullptr;
+        if (this_core >= 0) {
+            thread_on_core = scheduled_queue[this_core].front();
+        }
+        if (this_core < 0 || thread != thread_on_core) {
+            winner = thread;
+            break;
+        }
+        sug_cores.insert(this_core);
+    }
+    // if we got a suggested thread, select it, else do a second pass.
+    if (winner && winner->GetPriority() > 2) {
+        if (winner->IsRunning()) {
+            UnloadThread(static_cast<u32>(winner->GetProcessorID()));
+        }
+        TransferToCore(winner->GetPriority(), static_cast<s32>(core), winner);
+        update_thread(winner, sched);
+        return;
+    }
+    // Step 3: Select a suggested thread from another core
+    for (auto& src_core : sug_cores) {
+        auto it = scheduled_queue[src_core].begin();
+        it++;
+        if (it != scheduled_queue[src_core].end()) {
+            Thread* thread_on_core = scheduled_queue[src_core].front();
+            Thread* to_change = *it;
+            if (thread_on_core->IsRunning() || to_change->IsRunning()) {
+                UnloadThread(static_cast<u32>(src_core));
+            }
+            TransferToCore(thread_on_core->GetPriority(), static_cast<s32>(core), thread_on_core);
+            current_thread = thread_on_core;
+            break;
+        }
+    }
+    update_thread(current_thread, sched);
+}
+
+bool GlobalScheduler::YieldThread(Thread* yielding_thread) {
+    // Note: caller should use critical section, etc.
+    const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID());
+    const u32 priority = yielding_thread->GetPriority();
+
+    // Yield the thread
+    const Thread* const winner = scheduled_queue[core_id].front(priority);
+    ASSERT_MSG(yielding_thread == winner, "Thread yielding without being in front");
+    scheduled_queue[core_id].yield(priority);
+
+    return AskForReselectionOrMarkRedundant(yielding_thread, winner);
+}
+
+bool GlobalScheduler::YieldThreadAndBalanceLoad(Thread* yielding_thread) {
+    // Note: caller should check if !thread.IsSchedulerOperationRedundant and use critical section,
+    // etc.
+    const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID());
+    const u32 priority = yielding_thread->GetPriority();
+
+    // Yield the thread
+    ASSERT_MSG(yielding_thread == scheduled_queue[core_id].front(priority),
+               "Thread yielding without being in front");
+    scheduled_queue[core_id].yield(priority);
+
+    std::array<Thread*, NUM_CPU_CORES> current_threads;
+    for (u32 i = 0; i < NUM_CPU_CORES; i++) {
+        current_threads[i] = scheduled_queue[i].empty() ? nullptr : scheduled_queue[i].front();
+    }
+
+    Thread* next_thread = scheduled_queue[core_id].front(priority);
+    Thread* winner = nullptr;
+    for (auto& thread : suggested_queue[core_id]) {
+        const s32 source_core = thread->GetProcessorID();
+        if (source_core >= 0) {
+            if (current_threads[source_core] != nullptr) {
+                if (thread == current_threads[source_core] ||
+                    current_threads[source_core]->GetPriority() < min_regular_priority) {
+                    continue;
+                }
+            }
+        }
+        if (next_thread->GetLastRunningTicks() >= thread->GetLastRunningTicks() ||
+            next_thread->GetPriority() < thread->GetPriority()) {
+            if (thread->GetPriority() <= priority) {
+                winner = thread;
+                break;
+            }
+        }
+    }
+
+    if (winner != nullptr) {
+        if (winner != yielding_thread) {
+            if (winner->IsRunning()) {
+                UnloadThread(static_cast<u32>(winner->GetProcessorID()));
+            }
+            TransferToCore(winner->GetPriority(), s32(core_id), winner);
+        }
+    } else {
+        winner = next_thread;
+    }
+
+    return AskForReselectionOrMarkRedundant(yielding_thread, winner);
+}
+
+bool GlobalScheduler::YieldThreadAndWaitForLoadBalancing(Thread* yielding_thread) {
+    // Note: caller should check if !thread.IsSchedulerOperationRedundant and use critical section,
+    // etc.
+    Thread* winner = nullptr;
+    const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID());
+
+    // Remove the thread from its scheduled mlq, put it on the corresponding "suggested" one instead
+    TransferToCore(yielding_thread->GetPriority(), -1, yielding_thread);
+
+    // If the core is idle, perform load balancing, excluding the threads that have just used this
+    // function...
+    if (scheduled_queue[core_id].empty()) {
+        // Here, "current_threads" is calculated after the ""yield"", unlike yield -1
+        std::array<Thread*, NUM_CPU_CORES> current_threads;
+        for (u32 i = 0; i < NUM_CPU_CORES; i++) {
+            current_threads[i] = scheduled_queue[i].empty() ? nullptr : scheduled_queue[i].front();
+        }
+        for (auto& thread : suggested_queue[core_id]) {
+            const s32 source_core = thread->GetProcessorID();
+            if (source_core < 0 || thread == current_threads[source_core]) {
+                continue;
+            }
+            if (current_threads[source_core] == nullptr ||
+                current_threads[source_core]->GetPriority() >= min_regular_priority) {
+                winner = thread;
+            }
+            break;
+        }
+        if (winner != nullptr) {
+            if (winner != yielding_thread) {
+                if (winner->IsRunning()) {
+                    UnloadThread(static_cast<u32>(winner->GetProcessorID()));
+                }
+                TransferToCore(winner->GetPriority(), static_cast<s32>(core_id), winner);
+            }
+        } else {
+            winner = yielding_thread;
+        }
+    }
+
+    return AskForReselectionOrMarkRedundant(yielding_thread, winner);
+}
+
+void GlobalScheduler::PreemptThreads() {
+    for (std::size_t core_id = 0; core_id < NUM_CPU_CORES; core_id++) {
+        const u32 priority = preemption_priorities[core_id];
+
+        if (scheduled_queue[core_id].size(priority) > 0) {
+            scheduled_queue[core_id].front(priority)->IncrementYieldCount();
+            scheduled_queue[core_id].yield(priority);
+            if (scheduled_queue[core_id].size(priority) > 1) {
+                scheduled_queue[core_id].front(priority)->IncrementYieldCount();
+            }
+        }
+
+        Thread* current_thread =
+            scheduled_queue[core_id].empty() ? nullptr : scheduled_queue[core_id].front();
+        Thread* winner = nullptr;
+        for (auto& thread : suggested_queue[core_id]) {
+            const s32 source_core = thread->GetProcessorID();
+            if (thread->GetPriority() != priority) {
+                continue;
+            }
+            if (source_core >= 0) {
+                Thread* next_thread = scheduled_queue[source_core].empty()
+                                          ? nullptr
+                                          : scheduled_queue[source_core].front();
+                if (next_thread != nullptr && next_thread->GetPriority() < 2) {
+                    break;
+                }
+                if (next_thread == thread) {
+                    continue;
+                }
+            }
+            if (current_thread != nullptr &&
+                current_thread->GetLastRunningTicks() >= thread->GetLastRunningTicks()) {
+                winner = thread;
+                break;
+            }
+        }
+
+        if (winner != nullptr) {
+            if (winner->IsRunning()) {
+                UnloadThread(static_cast<u32>(winner->GetProcessorID()));
+            }
+            TransferToCore(winner->GetPriority(), s32(core_id), winner);
+            current_thread =
+                winner->GetPriority() <= current_thread->GetPriority() ? winner : current_thread;
+        }
+
+        if (current_thread != nullptr && current_thread->GetPriority() > priority) {
+            for (auto& thread : suggested_queue[core_id]) {
+                const s32 source_core = thread->GetProcessorID();
+                if (thread->GetPriority() < priority) {
+                    continue;
+                }
+                if (source_core >= 0) {
+                    Thread* next_thread = scheduled_queue[source_core].empty()
+                                              ? nullptr
+                                              : scheduled_queue[source_core].front();
+                    if (next_thread != nullptr && next_thread->GetPriority() < 2) {
+                        break;
+                    }
+                    if (next_thread == thread) {
+                        continue;
+                    }
+                }
+                if (current_thread != nullptr &&
+                    current_thread->GetLastRunningTicks() >= thread->GetLastRunningTicks()) {
+                    winner = thread;
+                    break;
+                }
+            }
+
+            if (winner != nullptr) {
+                if (winner->IsRunning()) {
+                    UnloadThread(static_cast<u32>(winner->GetProcessorID()));
+                }
+                TransferToCore(winner->GetPriority(), s32(core_id), winner);
+                current_thread = winner;
+            }
+        }
+
+        is_reselection_pending.store(true, std::memory_order_release);
     }
 }
 
+void GlobalScheduler::Suggest(u32 priority, std::size_t core, Thread* thread) {
+    suggested_queue[core].add(thread, priority);
+}
+
+void GlobalScheduler::Unsuggest(u32 priority, std::size_t core, Thread* thread) {
+    suggested_queue[core].remove(thread, priority);
+}
+
+void GlobalScheduler::Schedule(u32 priority, std::size_t core, Thread* thread) {
+    ASSERT_MSG(thread->GetProcessorID() == s32(core), "Thread must be assigned to this core.");
+    scheduled_queue[core].add(thread, priority);
+}
+
+void GlobalScheduler::SchedulePrepend(u32 priority, std::size_t core, Thread* thread) {
+    ASSERT_MSG(thread->GetProcessorID() == s32(core), "Thread must be assigned to this core.");
+    scheduled_queue[core].add(thread, priority, false);
+}
+
+void GlobalScheduler::Reschedule(u32 priority, std::size_t core, Thread* thread) {
+    scheduled_queue[core].remove(thread, priority);
+    scheduled_queue[core].add(thread, priority);
+}
+
+void GlobalScheduler::Unschedule(u32 priority, std::size_t core, Thread* thread) {
+    scheduled_queue[core].remove(thread, priority);
+}
+
+void GlobalScheduler::TransferToCore(u32 priority, s32 destination_core, Thread* thread) {
+    const bool schedulable = thread->GetPriority() < THREADPRIO_COUNT;
+    const s32 source_core = thread->GetProcessorID();
+    if (source_core == destination_core || !schedulable) {
+        return;
+    }
+    thread->SetProcessorID(destination_core);
+    if (source_core >= 0) {
+        Unschedule(priority, static_cast<u32>(source_core), thread);
+    }
+    if (destination_core >= 0) {
+        Unsuggest(priority, static_cast<u32>(destination_core), thread);
+        Schedule(priority, static_cast<u32>(destination_core), thread);
+    }
+    if (source_core >= 0) {
+        Suggest(priority, static_cast<u32>(source_core), thread);
+    }
+}
+
+bool GlobalScheduler::AskForReselectionOrMarkRedundant(Thread* current_thread,
+                                                       const Thread* winner) {
+    if (current_thread == winner) {
+        current_thread->IncrementYieldCount();
+        return true;
+    } else {
+        is_reselection_pending.store(true, std::memory_order_release);
+        return false;
+    }
+}
+
+void GlobalScheduler::Shutdown() {
+    for (std::size_t core = 0; core < NUM_CPU_CORES; core++) {
+        scheduled_queue[core].clear();
+        suggested_queue[core].clear();
+    }
+    thread_list.clear();
+}
+
+Scheduler::Scheduler(Core::System& system, Core::ARM_Interface& cpu_core, std::size_t core_id)
+    : system(system), cpu_core(cpu_core), core_id(core_id) {}
+
+Scheduler::~Scheduler() = default;
+
 bool Scheduler::HaveReadyThreads() const {
-    std::lock_guard lock{scheduler_mutex};
-    return !ready_queue.empty();
+    return system.GlobalScheduler().HaveReadyThreads(core_id);
 }
 
 Thread* Scheduler::GetCurrentThread() const {
     return current_thread.get();
 }
 
+Thread* Scheduler::GetSelectedThread() const {
+    return selected_thread.get();
+}
+
+void Scheduler::SelectThreads() {
+    system.GlobalScheduler().SelectThread(core_id);
+}
+
 u64 Scheduler::GetLastContextSwitchTicks() const {
     return last_context_switch_time;
 }
 
-Thread* Scheduler::PopNextReadyThread() {
-    Thread* next = nullptr;
-    Thread* thread = GetCurrentThread();
-
-    if (thread && thread->GetStatus() == ThreadStatus::Running) {
-        if (ready_queue.empty()) {
-            return thread;
-        }
-        // We have to do better than the current thread.
-        // This call returns null when that's not possible.
-        next = ready_queue.front();
-        if (next == nullptr || next->GetPriority() >= thread->GetPriority()) {
-            next = thread;
-        }
-    } else {
-        if (ready_queue.empty()) {
-            return nullptr;
-        }
-        next = ready_queue.front();
+void Scheduler::TryDoContextSwitch() {
+    if (is_context_switch_pending) {
+        SwitchContext();
     }
-
-    return next;
 }
 
-void Scheduler::SwitchContext(Thread* new_thread) {
-    Thread* previous_thread = GetCurrentThread();
+void Scheduler::UnloadThread() {
+    Thread* const previous_thread = GetCurrentThread();
     Process* const previous_process = system.Kernel().CurrentProcess();
 
     UpdateLastContextSwitchTime(previous_thread, previous_process);
@@ -80,23 +403,52 @@ void Scheduler::SwitchContext(Thread* new_thread) {
         if (previous_thread->GetStatus() == ThreadStatus::Running) {
             // This is only the case when a reschedule is triggered without the current thread
             // yielding execution (i.e. an event triggered, system core time-sliced, etc)
-            ready_queue.add(previous_thread, previous_thread->GetPriority(), false);
             previous_thread->SetStatus(ThreadStatus::Ready);
         }
+        previous_thread->SetIsRunning(false);
+    }
+    current_thread = nullptr;
+}
+
+void Scheduler::SwitchContext() {
+    Thread* const previous_thread = GetCurrentThread();
+    Thread* const new_thread = GetSelectedThread();
+
+    is_context_switch_pending = false;
+    if (new_thread == previous_thread) {
+        return;
+    }
+
+    Process* const previous_process = system.Kernel().CurrentProcess();
+
+    UpdateLastContextSwitchTime(previous_thread, previous_process);
+
+    // Save context for previous thread
+    if (previous_thread) {
+        cpu_core.SaveContext(previous_thread->GetContext());
+        // Save the TPIDR_EL0 system register in case it was modified.
+        previous_thread->SetTPIDR_EL0(cpu_core.GetTPIDR_EL0());
+
+        if (previous_thread->GetStatus() == ThreadStatus::Running) {
+            // This is only the case when a reschedule is triggered without the current thread
+            // yielding execution (i.e. an event triggered, system core time-sliced, etc)
+            previous_thread->SetStatus(ThreadStatus::Ready);
+        }
+        previous_thread->SetIsRunning(false);
     }
 
     // Load context of new thread
     if (new_thread) {
+        ASSERT_MSG(new_thread->GetProcessorID() == s32(this->core_id),
+                   "Thread must be assigned to this core.");
         ASSERT_MSG(new_thread->GetStatus() == ThreadStatus::Ready,
                    "Thread must be ready to become running.");
 
         // Cancel any outstanding wakeup events for this thread
         new_thread->CancelWakeupTimer();
-
         current_thread = new_thread;
-
-        ready_queue.remove(new_thread, new_thread->GetPriority());
         new_thread->SetStatus(ThreadStatus::Running);
+        new_thread->SetIsRunning(true);
 
         auto* const thread_owner_process = current_thread->GetOwnerProcess();
         if (previous_process != thread_owner_process) {
@@ -130,124 +482,9 @@ void Scheduler::UpdateLastContextSwitchTime(Thread* thread, Process* process) {
     last_context_switch_time = most_recent_switch_ticks;
 }
 
-void Scheduler::Reschedule() {
-    std::lock_guard lock{scheduler_mutex};
-
-    Thread* cur = GetCurrentThread();
-    Thread* next = PopNextReadyThread();
-
-    if (cur && next) {
-        LOG_TRACE(Kernel, "context switch {} -> {}", cur->GetObjectId(), next->GetObjectId());
-    } else if (cur) {
-        LOG_TRACE(Kernel, "context switch {} -> idle", cur->GetObjectId());
-    } else if (next) {
-        LOG_TRACE(Kernel, "context switch idle -> {}", next->GetObjectId());
-    }
-
-    SwitchContext(next);
-}
-
-void Scheduler::AddThread(SharedPtr<Thread> thread) {
-    std::lock_guard lock{scheduler_mutex};
-
-    thread_list.push_back(std::move(thread));
-}
-
-void Scheduler::RemoveThread(Thread* thread) {
-    std::lock_guard lock{scheduler_mutex};
-
-    thread_list.erase(std::remove(thread_list.begin(), thread_list.end(), thread),
-                      thread_list.end());
-}
-
-void Scheduler::ScheduleThread(Thread* thread, u32 priority) {
-    std::lock_guard lock{scheduler_mutex};
-
-    ASSERT(thread->GetStatus() == ThreadStatus::Ready);
-    ready_queue.add(thread, priority);
-}
-
-void Scheduler::UnscheduleThread(Thread* thread, u32 priority) {
-    std::lock_guard lock{scheduler_mutex};
-
-    ASSERT(thread->GetStatus() == ThreadStatus::Ready);
-    ready_queue.remove(thread, priority);
-}
-
-void Scheduler::SetThreadPriority(Thread* thread, u32 priority) {
-    std::lock_guard lock{scheduler_mutex};
-    if (thread->GetPriority() == priority) {
-        return;
-    }
-
-    // If thread was ready, adjust queues
-    if (thread->GetStatus() == ThreadStatus::Ready)
-        ready_queue.adjust(thread, thread->GetPriority(), priority);
-}
-
-Thread* Scheduler::GetNextSuggestedThread(u32 core, u32 maximum_priority) const {
-    std::lock_guard lock{scheduler_mutex};
-
-    const u32 mask = 1U << core;
-    for (auto* thread : ready_queue) {
-        if ((thread->GetAffinityMask() & mask) != 0 && thread->GetPriority() < maximum_priority) {
-            return thread;
-        }
-    }
-    return nullptr;
-}
-
-void Scheduler::YieldWithoutLoadBalancing(Thread* thread) {
-    ASSERT(thread != nullptr);
-    // Avoid yielding if the thread isn't even running.
-    ASSERT(thread->GetStatus() == ThreadStatus::Running);
-
-    // Sanity check that the priority is valid
-    ASSERT(thread->GetPriority() < THREADPRIO_COUNT);
-
-    // Yield this thread -- sleep for zero time and force reschedule to different thread
-    GetCurrentThread()->Sleep(0);
-}
-
-void Scheduler::YieldWithLoadBalancing(Thread* thread) {
-    ASSERT(thread != nullptr);
-    const auto priority = thread->GetPriority();
-    const auto core = static_cast<u32>(thread->GetProcessorID());
-
-    // Avoid yielding if the thread isn't even running.
-    ASSERT(thread->GetStatus() == ThreadStatus::Running);
-
-    // Sanity check that the priority is valid
-    ASSERT(priority < THREADPRIO_COUNT);
-
-    // Sleep for zero time to be able to force reschedule to different thread
-    GetCurrentThread()->Sleep(0);
-
-    Thread* suggested_thread = nullptr;
-
-    // Search through all of the cpu cores (except this one) for a suggested thread.
-    // Take the first non-nullptr one
-    for (unsigned cur_core = 0; cur_core < Core::NUM_CPU_CORES; ++cur_core) {
-        const auto res =
-            system.CpuCore(cur_core).Scheduler().GetNextSuggestedThread(core, priority);
-
-        // If scheduler provides a suggested thread
-        if (res != nullptr) {
-            // And its better than the current suggested thread (or is the first valid one)
-            if (suggested_thread == nullptr ||
-                suggested_thread->GetPriority() > res->GetPriority()) {
-                suggested_thread = res;
-            }
-        }
-    }
-
-    // If a suggested thread was found, queue that for this core
-    if (suggested_thread != nullptr)
-        suggested_thread->ChangeCore(core, suggested_thread->GetAffinityMask());
-}
-
-void Scheduler::YieldAndWaitForLoadBalancing(Thread* thread) {
-    UNIMPLEMENTED_MSG("Wait for load balancing thread yield type is not implemented!");
+void Scheduler::Shutdown() {
+    current_thread = nullptr;
+    selected_thread = nullptr;
 }
 
 } // namespace Kernel

src/core/hle/kernel/scheduler.h

@@ -20,124 +20,185 @@ namespace Kernel {
 
 class Process;
 
-class Scheduler final {
+class GlobalScheduler final {
 public:
-    explicit Scheduler(Core::System& system, Core::ARM_Interface& cpu_core);
-    ~Scheduler();
+    static constexpr u32 NUM_CPU_CORES = 4;
 
-    /// Returns whether there are any threads that are ready to run.
-    bool HaveReadyThreads() const;
-
-    /// Reschedules to the next available thread (call after current thread is suspended)
-    void Reschedule();
-
-    /// Gets the current running thread
-    Thread* GetCurrentThread() const;
-
-    /// Gets the timestamp for the last context switch in ticks.
-    u64 GetLastContextSwitchTicks() const;
+    explicit GlobalScheduler(Core::System& system);
+    ~GlobalScheduler();
 
     /// Adds a new thread to the scheduler
     void AddThread(SharedPtr<Thread> thread);
 
     /// Removes a thread from the scheduler
-    void RemoveThread(Thread* thread);
-
-    /// Schedules a thread that has become "ready"
-    void ScheduleThread(Thread* thread, u32 priority);
-
-    /// Unschedules a thread that was already scheduled
-    void UnscheduleThread(Thread* thread, u32 priority);
-
-    /// Sets the priority of a thread in the scheduler
-    void SetThreadPriority(Thread* thread, u32 priority);
-
-    /// Gets the next suggested thread for load balancing
-    Thread* GetNextSuggestedThread(u32 core, u32 minimum_priority) const;
-
-    /**
-     * YieldWithoutLoadBalancing -- analogous to normal yield on a system
-     * Moves the thread to the end of the ready queue for its priority, and then reschedules the
-     * system to the new head of the queue.
-     *
-     * Example (Single Core -- but can be extrapolated to multi):
-     * ready_queue[prio=0]: ThreadA, ThreadB, ThreadC (->exec order->)
-     * Currently Running: ThreadR
-     *
-     * ThreadR calls YieldWithoutLoadBalancing
-     *
-     * ThreadR is moved to the end of ready_queue[prio=0]:
-     * ready_queue[prio=0]: ThreadA, ThreadB, ThreadC, ThreadR (->exec order->)
-     * Currently Running: Nothing
-     *
-     * System is rescheduled (ThreadA is popped off of queue):
-     * ready_queue[prio=0]: ThreadB, ThreadC, ThreadR (->exec order->)
-     * Currently Running: ThreadA
-     *
-     * If the queue is empty at time of call, no yielding occurs. This does not cross between cores
-     * or priorities at all.
-     */
-    void YieldWithoutLoadBalancing(Thread* thread);
-
-    /**
-     * YieldWithLoadBalancing -- yield but with better selection of the new running thread
-     * Moves the current thread to the end of the ready queue for its priority, then selects a
-     * 'suggested thread' (a thread on a different core that could run on this core) from the
-     * scheduler, changes its core, and reschedules the current core to that thread.
-     *
-     * Example (Dual Core -- can be extrapolated to Quad Core, this is just normal yield if it were
-     * single core):
-     * ready_queue[core=0][prio=0]: ThreadA, ThreadB (affinities not pictured as irrelevant
-     * ready_queue[core=1][prio=0]: ThreadC[affinity=both], ThreadD[affinity=core1only]
-     * Currently Running: ThreadQ on Core 0 || ThreadP on Core 1
-     *
-     * ThreadQ calls YieldWithLoadBalancing
-     *
-     * ThreadQ is moved to the end of ready_queue[core=0][prio=0]:
-     * ready_queue[core=0][prio=0]: ThreadA, ThreadB
-     * ready_queue[core=1][prio=0]: ThreadC[affinity=both], ThreadD[affinity=core1only]
-     * Currently Running: ThreadQ on Core 0 || ThreadP on Core 1
-     *
-     * A list of suggested threads for each core is compiled
-     * Suggested Threads: {ThreadC on Core 1}
-     * If this were quad core (as the switch is), there could be between 0 and 3 threads in this
-     * list. If there are more than one, the thread is selected by highest prio.
-     *
-     * ThreadC is core changed to Core 0:
-     * ready_queue[core=0][prio=0]: ThreadC, ThreadA, ThreadB, ThreadQ
-     * ready_queue[core=1][prio=0]: ThreadD
-     * Currently Running: None on Core 0 || ThreadP on Core 1
-     *
-     * System is rescheduled (ThreadC is popped off of queue):
-     * ready_queue[core=0][prio=0]: ThreadA, ThreadB, ThreadQ
-     * ready_queue[core=1][prio=0]: ThreadD
-     * Currently Running: ThreadC on Core 0 || ThreadP on Core 1
-     *
-     * If no suggested threads can be found this will behave just as normal yield. If there are
-     * multiple candidates for the suggested thread on a core, the highest prio is taken.
-     */
-    void YieldWithLoadBalancing(Thread* thread);
-
-    /// Currently unknown -- asserts as unimplemented on call
-    void YieldAndWaitForLoadBalancing(Thread* thread);
+    void RemoveThread(const Thread* thread);
 
     /// Returns a list of all threads managed by the scheduler
     const std::vector<SharedPtr<Thread>>& GetThreadList() const {
         return thread_list;
     }
 
-private:
     /**
-     * Pops and returns the next thread from the thread queue
-     * @return A pointer to the next ready thread
+     * Add a thread to the suggested queue of a cpu core. Suggested threads may be
+     * picked if no thread is scheduled to run on the core.
      */
-    Thread* PopNextReadyThread();
+    void Suggest(u32 priority, std::size_t core, Thread* thread);
 
     /**
-     * Switches the CPU's active thread context to that of the specified thread
-     * @param new_thread The thread to switch to
+     * Remove a thread to the suggested queue of a cpu core. Suggested threads may be
+     * picked if no thread is scheduled to run on the core.
      */
-    void SwitchContext(Thread* new_thread);
+    void Unsuggest(u32 priority, std::size_t core, Thread* thread);
+
+    /**
+     * Add a thread to the scheduling queue of a cpu core. The thread is added at the
+     * back the queue in its priority level.
+     */
+    void Schedule(u32 priority, std::size_t core, Thread* thread);
+
+    /**
+     * Add a thread to the scheduling queue of a cpu core. The thread is added at the
+     * front the queue in its priority level.
+     */
+    void SchedulePrepend(u32 priority, std::size_t core, Thread* thread);
+
+    /// Reschedule an already scheduled thread based on a new priority
+    void Reschedule(u32 priority, std::size_t core, Thread* thread);
+
+    /// Unschedules a thread.
+    void Unschedule(u32 priority, std::size_t core, Thread* thread);
+
+    /// Selects a core and forces it to unload its current thread's context
+    void UnloadThread(std::size_t core);
+
+    /**
+     * Takes care of selecting the new scheduled thread in three steps:
+     *
+     * 1. First a thread is selected from the top of the priority queue. If no thread
+     *    is obtained then we move to step two, else we are done.
+     *
+     * 2. Second we try to get a suggested thread that's not assigned to any core or
+     *    that is not the top thread in that core.
+     *
+     * 3. Third is no suggested thread is found, we do a second pass and pick a running
+     *    thread in another core and swap it with its current thread.
+     */
+    void SelectThread(std::size_t core);
+
+    bool HaveReadyThreads(std::size_t core_id) const {
+        return !scheduled_queue[core_id].empty();
+    }
+
+    /**
+     * Takes a thread and moves it to the back of the it's priority list.
+     *
+     * @note This operation can be redundant and no scheduling is changed if marked as so.
+     */
+    bool YieldThread(Thread* thread);
+
+    /**
+     * Takes a thread and moves it to the back of the it's priority list.
+     * Afterwards, tries to pick a suggested thread from the suggested queue that has worse time or
+     * a better priority than the next thread in the core.
+     *
+     * @note This operation can be redundant and no scheduling is changed if marked as so.
+     */
+    bool YieldThreadAndBalanceLoad(Thread* thread);
+
+    /**
+     * Takes a thread and moves it out of the scheduling queue.
+     * and into the suggested queue. If no thread can be scheduled afterwards in that core,
+     * a suggested thread is obtained instead.
+     *
+     * @note This operation can be redundant and no scheduling is changed if marked as so.
+     */
+    bool YieldThreadAndWaitForLoadBalancing(Thread* thread);
+
+    /**
+     * Rotates the scheduling queues of threads at a preemption priority and then does
+     * some core rebalancing. Preemption priorities can be found in the array
+     * 'preemption_priorities'.
+     *
+     * @note This operation happens every 10ms.
+     */
+    void PreemptThreads();
+
+    u32 CpuCoresCount() const {
+        return NUM_CPU_CORES;
+    }
+
+    void SetReselectionPending() {
+        is_reselection_pending.store(true, std::memory_order_release);
+    }
+
+    bool IsReselectionPending() const {
+        return is_reselection_pending.load(std::memory_order_acquire);
+    }
+
+    void Shutdown();
+
+private:
+    /**
+     * Transfers a thread into an specific core. If the destination_core is -1
+     * it will be unscheduled from its source code and added into its suggested
+     * queue.
+     */
+    void TransferToCore(u32 priority, s32 destination_core, Thread* thread);
+
+    bool AskForReselectionOrMarkRedundant(Thread* current_thread, const Thread* winner);
+
+    static constexpr u32 min_regular_priority = 2;
+    std::array<Common::MultiLevelQueue<Thread*, THREADPRIO_COUNT>, NUM_CPU_CORES> scheduled_queue;
+    std::array<Common::MultiLevelQueue<Thread*, THREADPRIO_COUNT>, NUM_CPU_CORES> suggested_queue;
+    std::atomic<bool> is_reselection_pending{false};
+
+    // The priority levels at which the global scheduler preempts threads every 10 ms. They are
+    // ordered from Core 0 to Core 3.
+    std::array<u32, NUM_CPU_CORES> preemption_priorities = {59, 59, 59, 62};
+
+    /// Lists all thread ids that aren't deleted/etc.
+    std::vector<SharedPtr<Thread>> thread_list;
+    Core::System& system;
+};
+
+class Scheduler final {
+public:
+    explicit Scheduler(Core::System& system, Core::ARM_Interface& cpu_core, std::size_t core_id);
+    ~Scheduler();
+
+    /// Returns whether there are any threads that are ready to run.
+    bool HaveReadyThreads() const;
+
+    /// Reschedules to the next available thread (call after current thread is suspended)
+    void TryDoContextSwitch();
+
+    /// Unloads currently running thread
+    void UnloadThread();
+
+    /// Select the threads in top of the scheduling multilist.
+    void SelectThreads();
+
+    /// Gets the current running thread
+    Thread* GetCurrentThread() const;
+
+    /// Gets the currently selected thread from the top of the multilevel queue
+    Thread* GetSelectedThread() const;
+
+    /// Gets the timestamp for the last context switch in ticks.
+    u64 GetLastContextSwitchTicks() const;
+
+    bool ContextSwitchPending() const {
+        return is_context_switch_pending;
+    }
+
+    /// Shutdowns the scheduler.
+    void Shutdown();
+
+private:
+    friend class GlobalScheduler;
+
+    /// Switches the CPU's active thread context to that of the specified thread
+    void SwitchContext();
 
     /**
      * Called on every context switch to update the internal timestamp
@@ -152,19 +213,16 @@ private:
      */
     void UpdateLastContextSwitchTime(Thread* thread, Process* process);
 
-    /// Lists all thread ids that aren't deleted/etc.
-    std::vector<SharedPtr<Thread>> thread_list;
-
-    /// Lists only ready thread ids.
-    Common::MultiLevelQueue<Thread*, THREADPRIO_LOWEST + 1> ready_queue;
-
     SharedPtr<Thread> current_thread = nullptr;
-
-    Core::ARM_Interface& cpu_core;
-    u64 last_context_switch_time = 0;
+    SharedPtr<Thread> selected_thread = nullptr;
 
     Core::System& system;
-    static std::mutex scheduler_mutex;
+    Core::ARM_Interface& cpu_core;
+    u64 last_context_switch_time = 0;
+    u64 idle_selection_count = 0;
+    const std::size_t core_id;
+
+    bool is_context_switch_pending = false;
 };
 
 } // namespace Kernel

src/core/hle/kernel/svc.cpp

@@ -516,7 +516,7 @@ static ResultCode WaitSynchronization(Core::System& system, Handle* index, VAddr
     thread->WakeAfterDelay(nano_seconds);
     thread->SetWakeupCallback(DefaultThreadWakeupCallback);
 
-    system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(thread->GetProcessorID());
 
     return RESULT_TIMEOUT;
 }
@@ -534,6 +534,7 @@ static ResultCode CancelSynchronization(Core::System& system, Handle thread_hand
     }
 
     thread->CancelWait();
+    system.PrepareReschedule(thread->GetProcessorID());
     return RESULT_SUCCESS;
 }
 
@@ -1066,6 +1067,8 @@ static ResultCode SetThreadActivity(Core::System& system, Handle handle, u32 act
     }
 
     thread->SetActivity(static_cast<ThreadActivity>(activity));
+
+    system.PrepareReschedule(thread->GetProcessorID());
     return RESULT_SUCCESS;
 }
 
@@ -1147,7 +1150,7 @@ static ResultCode SetThreadPriority(Core::System& system, Handle handle, u32 pri
 
     thread->SetPriority(priority);
 
-    system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(thread->GetProcessorID());
     return RESULT_SUCCESS;
 }
 
@@ -1503,7 +1506,7 @@ static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr e
     thread->SetName(
         fmt::format("thread[entry_point={:X}, handle={:X}]", entry_point, *new_thread_handle));
 
-    system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(thread->GetProcessorID());
 
     return RESULT_SUCCESS;
 }
@@ -1525,7 +1528,7 @@ static ResultCode StartThread(Core::System& system, Handle thread_handle) {
     thread->ResumeFromWait();
 
     if (thread->GetStatus() == ThreadStatus::Ready) {
-        system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+        system.PrepareReschedule(thread->GetProcessorID());
     }
 
     return RESULT_SUCCESS;
@@ -1537,7 +1540,7 @@ static void ExitThread(Core::System& system) {
 
     auto* const current_thread = system.CurrentScheduler().GetCurrentThread();
     current_thread->Stop();
-    system.CurrentScheduler().RemoveThread(current_thread);
+    system.GlobalScheduler().RemoveThread(current_thread);
     system.PrepareReschedule();
 }
 
@@ -1553,17 +1556,18 @@ static void SleepThread(Core::System& system, s64 nanoseconds) {
 
     auto& scheduler = system.CurrentScheduler();
     auto* const current_thread = scheduler.GetCurrentThread();
+    bool is_redundant = false;
 
     if (nanoseconds <= 0) {
         switch (static_cast<SleepType>(nanoseconds)) {
         case SleepType::YieldWithoutLoadBalancing:
-            scheduler.YieldWithoutLoadBalancing(current_thread);
+            is_redundant = current_thread->YieldSimple();
             break;
         case SleepType::YieldWithLoadBalancing:
-            scheduler.YieldWithLoadBalancing(current_thread);
+            is_redundant = current_thread->YieldAndBalanceLoad();
             break;
         case SleepType::YieldAndWaitForLoadBalancing:
-            scheduler.YieldAndWaitForLoadBalancing(current_thread);
+            is_redundant = current_thread->YieldAndWaitForLoadBalancing();
             break;
         default:
             UNREACHABLE_MSG("Unimplemented sleep yield type '{:016X}'!", nanoseconds);
@@ -1572,10 +1576,13 @@ static void SleepThread(Core::System& system, s64 nanoseconds) {
         current_thread->Sleep(nanoseconds);
     }
 
-    // Reschedule all CPU cores
-    for (std::size_t i = 0; i < Core::NUM_CPU_CORES; ++i) {
-        system.CpuCore(i).PrepareReschedule();
+    if (is_redundant) {
+        // If it's redundant, the core is pretty much idle. Some games keep idling
+        // a core while it's doing nothing, we advance timing to avoid costly continuous
+        // calls.
+        system.CoreTiming().AddTicks(2000);
     }
+    system.PrepareReschedule(current_thread->GetProcessorID());
 }
 
 /// Wait process wide key atomic
@@ -1601,6 +1608,8 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add
         return ERR_INVALID_ADDRESS;
     }
 
+    ASSERT(condition_variable_addr == Common::AlignDown(condition_variable_addr, 4));
+
     auto* const current_process = system.Kernel().CurrentProcess();
     const auto& handle_table = current_process->GetHandleTable();
     SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
@@ -1622,7 +1631,7 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add
 
     // Note: Deliberately don't attempt to inherit the lock owner's priority.
 
-    system.CpuCore(current_thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(current_thread->GetProcessorID());
     return RESULT_SUCCESS;
 }
 
@@ -1632,24 +1641,19 @@ static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_var
     LOG_TRACE(Kernel_SVC, "called, condition_variable_addr=0x{:X}, target=0x{:08X}",
               condition_variable_addr, target);
 
-    const auto RetrieveWaitingThreads = [&system](std::size_t core_index,
-                                                  std::vector<SharedPtr<Thread>>& waiting_threads,
-                                                  VAddr condvar_addr) {
-        const auto& scheduler = system.Scheduler(core_index);
-        const auto& thread_list = scheduler.GetThreadList();
-
-        for (const auto& thread : thread_list) {
-            if (thread->GetCondVarWaitAddress() == condvar_addr)
-                waiting_threads.push_back(thread);
-        }
-    };
+    ASSERT(condition_variable_addr == Common::AlignDown(condition_variable_addr, 4));
 
     // Retrieve a list of all threads that are waiting for this condition variable.
     std::vector<SharedPtr<Thread>> waiting_threads;
-    RetrieveWaitingThreads(0, waiting_threads, condition_variable_addr);
-    RetrieveWaitingThreads(1, waiting_threads, condition_variable_addr);
-    RetrieveWaitingThreads(2, waiting_threads, condition_variable_addr);
-    RetrieveWaitingThreads(3, waiting_threads, condition_variable_addr);
+    const auto& scheduler = system.GlobalScheduler();
+    const auto& thread_list = scheduler.GetThreadList();
+
+    for (const auto& thread : thread_list) {
+        if (thread->GetCondVarWaitAddress() == condition_variable_addr) {
+            waiting_threads.push_back(thread);
+        }
+    }
+
     // Sort them by priority, such that the highest priority ones come first.
     std::sort(waiting_threads.begin(), waiting_threads.end(),
               [](const SharedPtr<Thread>& lhs, const SharedPtr<Thread>& rhs) {
@@ -1679,18 +1683,20 @@ static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_var
 
         // Atomically read the value of the mutex.
         u32 mutex_val = 0;
+        u32 update_val = 0;
+        const VAddr mutex_address = thread->GetMutexWaitAddress();
         do {
-            monitor.SetExclusive(current_core, thread->GetMutexWaitAddress());
+            monitor.SetExclusive(current_core, mutex_address);
 
             // If the mutex is not yet acquired, acquire it.
-            mutex_val = Memory::Read32(thread->GetMutexWaitAddress());
+            mutex_val = Memory::Read32(mutex_address);
 
             if (mutex_val != 0) {
-                monitor.ClearExclusive();
-                break;
+                update_val = mutex_val | Mutex::MutexHasWaitersFlag;
+            } else {
+                update_val = thread->GetWaitHandle();
             }
-        } while (!monitor.ExclusiveWrite32(current_core, thread->GetMutexWaitAddress(),
-                                           thread->GetWaitHandle()));
+        } while (!monitor.ExclusiveWrite32(current_core, mutex_address, update_val));
         if (mutex_val == 0) {
             // We were able to acquire the mutex, resume this thread.
             ASSERT(thread->GetStatus() == ThreadStatus::WaitCondVar);
@@ -1704,20 +1710,9 @@ static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_var
             thread->SetLockOwner(nullptr);
             thread->SetMutexWaitAddress(0);
             thread->SetWaitHandle(0);
-            system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+            thread->SetWaitSynchronizationResult(RESULT_SUCCESS);
+            system.PrepareReschedule(thread->GetProcessorID());
         } else {
-            // Atomically signal that the mutex now has a waiting thread.
-            do {
-                monitor.SetExclusive(current_core, thread->GetMutexWaitAddress());
-
-                // Ensure that the mutex value is still what we expect.
-                u32 value = Memory::Read32(thread->GetMutexWaitAddress());
-                // TODO(Subv): When this happens, the kernel just clears the exclusive state and
-                // retries the initial read for this thread.
-                ASSERT_MSG(mutex_val == value, "Unhandled synchronization primitive case");
-            } while (!monitor.ExclusiveWrite32(current_core, thread->GetMutexWaitAddress(),
-                                               mutex_val | Mutex::MutexHasWaitersFlag));
-
             // The mutex is already owned by some other thread, make this thread wait on it.
             const Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask);
             const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
@@ -1728,6 +1723,7 @@ static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_var
             thread->SetStatus(ThreadStatus::WaitMutex);
 
             owner->AddMutexWaiter(thread);
+            system.PrepareReschedule(thread->GetProcessorID());
         }
     }
 
@@ -1754,7 +1750,12 @@ static ResultCode WaitForAddress(Core::System& system, VAddr address, u32 type,
 
     const auto arbitration_type = static_cast<AddressArbiter::ArbitrationType>(type);
     auto& address_arbiter = system.Kernel().CurrentProcess()->GetAddressArbiter();
-    return address_arbiter.WaitForAddress(address, arbitration_type, value, timeout);
+    const ResultCode result =
+        address_arbiter.WaitForAddress(address, arbitration_type, value, timeout);
+    if (result == RESULT_SUCCESS) {
+        system.PrepareReschedule();
+    }
+    return result;
 }
 
 // Signals to an address (via Address Arbiter)
@@ -2040,7 +2041,10 @@ static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle,
         return ERR_INVALID_HANDLE;
     }
 
+    system.PrepareReschedule(thread->GetProcessorID());
     thread->ChangeCore(core, affinity_mask);
+    system.PrepareReschedule(thread->GetProcessorID());
+
     return RESULT_SUCCESS;
 }
 
@@ -2095,7 +2099,7 @@ static ResultCode CreateEvent(Core::System& system, Handle* write_handle, Handle
 
     auto& kernel = system.Kernel();
     const auto [readable_event, writable_event] =
-        WritableEvent::CreateEventPair(kernel, ResetType::Manual, "CreateEvent");
+        WritableEvent::CreateEventPair(kernel, "CreateEvent");
 
     HandleTable& handle_table = kernel.CurrentProcess()->GetHandleTable();
 
@@ -2151,6 +2155,7 @@ static ResultCode SignalEvent(Core::System& system, Handle handle) {
     }
 
     writable_event->Signal();
+    system.PrepareReschedule();
     return RESULT_SUCCESS;
 }
 

src/core/hle/kernel/thread.cpp

@@ -45,15 +45,7 @@ void Thread::Stop() {
                                                              callback_handle);
     kernel.ThreadWakeupCallbackHandleTable().Close(callback_handle);
     callback_handle = 0;
-
-    // Clean up thread from ready queue
-    // This is only needed when the thread is terminated forcefully (SVC TerminateProcess)
-    if (status == ThreadStatus::Ready || status == ThreadStatus::Paused) {
-        scheduler->UnscheduleThread(this, current_priority);
-    }
-
-    status = ThreadStatus::Dead;
-
+    SetStatus(ThreadStatus::Dead);
     WakeupAllWaitingThreads();
 
     // Clean up any dangling references in objects that this thread was waiting for
@@ -85,18 +77,6 @@ void Thread::CancelWakeupTimer() {
                                                              callback_handle);
 }
 
-static std::optional<s32> GetNextProcessorId(u64 mask) {
-    for (s32 index = 0; index < Core::NUM_CPU_CORES; ++index) {
-        if (mask & (1ULL << index)) {
-            if (!Core::System::GetInstance().Scheduler(index).GetCurrentThread()) {
-                // Core is enabled and not running any threads, use this one
-                return index;
-            }
-        }
-    }
-    return {};
-}
-
 void Thread::ResumeFromWait() {
     ASSERT_MSG(wait_objects.empty(), "Thread is waking up while waiting for objects");
 
@@ -132,17 +112,16 @@ void Thread::ResumeFromWait() {
     wakeup_callback = nullptr;
 
     if (activity == ThreadActivity::Paused) {
-        status = ThreadStatus::Paused;
+        SetStatus(ThreadStatus::Paused);
         return;
     }
 
-    status = ThreadStatus::Ready;
-
-    ChangeScheduler();
+    SetStatus(ThreadStatus::Ready);
 }
 
 void Thread::CancelWait() {
     ASSERT(GetStatus() == ThreadStatus::WaitSynch);
+    ClearWaitObjects();
     SetWaitSynchronizationResult(ERR_SYNCHRONIZATION_CANCELED);
     ResumeFromWait();
 }
@@ -182,7 +161,7 @@ ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name
     if (!Memory::IsValidVirtualAddress(owner_process, entry_point)) {
         LOG_ERROR(Kernel_SVC, "(name={}): invalid entry {:016X}", name, entry_point);
         // TODO (bunnei): Find the correct error code to use here
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
 
     auto& system = Core::System::GetInstance();
@@ -205,9 +184,9 @@ ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name
     thread->name = std::move(name);
     thread->callback_handle = kernel.ThreadWakeupCallbackHandleTable().Create(thread).Unwrap();
     thread->owner_process = &owner_process;
+    auto& scheduler = kernel.GlobalScheduler();
+    scheduler.AddThread(thread);
     thread->tls_address = thread->owner_process->CreateTLSRegion();
-    thread->scheduler = &system.Scheduler(processor_id);
-    thread->scheduler->AddThread(thread);
 
     thread->owner_process->RegisterThread(thread.get());
 
@@ -250,6 +229,22 @@ void Thread::SetStatus(ThreadStatus new_status) {
         return;
     }
 
+    switch (new_status) {
+    case ThreadStatus::Ready:
+    case ThreadStatus::Running:
+        SetSchedulingStatus(ThreadSchedStatus::Runnable);
+        break;
+    case ThreadStatus::Dormant:
+        SetSchedulingStatus(ThreadSchedStatus::None);
+        break;
+    case ThreadStatus::Dead:
+        SetSchedulingStatus(ThreadSchedStatus::Exited);
+        break;
+    default:
+        SetSchedulingStatus(ThreadSchedStatus::Paused);
+        break;
+    }
+
     if (status == ThreadStatus::Running) {
         last_running_ticks = Core::System::GetInstance().CoreTiming().GetTicks();
     }
@@ -311,8 +306,7 @@ void Thread::UpdatePriority() {
         return;
     }
 
-    scheduler->SetThreadPriority(this, new_priority);
-    current_priority = new_priority;
+    SetCurrentPriority(new_priority);
 
     if (!lock_owner) {
         return;
@@ -328,47 +322,7 @@ void Thread::UpdatePriority() {
 }
 
 void Thread::ChangeCore(u32 core, u64 mask) {
-    ideal_core = core;
-    affinity_mask = mask;
-    ChangeScheduler();
-}
-
-void Thread::ChangeScheduler() {
-    if (status != ThreadStatus::Ready) {
-        return;
-    }
-
-    auto& system = Core::System::GetInstance();
-    std::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)};
-
-    if (!new_processor_id) {
-        new_processor_id = processor_id;
-    }
-    if (ideal_core != -1 && system.Scheduler(ideal_core).GetCurrentThread() == nullptr) {
-        new_processor_id = ideal_core;
-    }
-
-    ASSERT(*new_processor_id < 4);
-
-    // Add thread to new core's scheduler
-    auto& next_scheduler = system.Scheduler(*new_processor_id);
-
-    if (*new_processor_id != processor_id) {
-        // Remove thread from previous core's scheduler
-        scheduler->RemoveThread(this);
-        next_scheduler.AddThread(this);
-    }
-
-    processor_id = *new_processor_id;
-
-    // If the thread was ready, unschedule from the previous core and schedule on the new core
-    scheduler->UnscheduleThread(this, current_priority);
-    next_scheduler.ScheduleThread(this, current_priority);
-
-    // Change thread's scheduler
-    scheduler = &next_scheduler;
-
-    system.CpuCore(processor_id).PrepareReschedule();
+    SetCoreAndAffinityMask(core, mask);
 }
 
 bool Thread::AllWaitObjectsReady() const {
@@ -388,10 +342,8 @@ void Thread::SetActivity(ThreadActivity value) {
 
     if (value == ThreadActivity::Paused) {
         // Set status if not waiting
-        if (status == ThreadStatus::Ready) {
-            status = ThreadStatus::Paused;
-        } else if (status == ThreadStatus::Running) {
-            status = ThreadStatus::Paused;
+        if (status == ThreadStatus::Ready || status == ThreadStatus::Running) {
+            SetStatus(ThreadStatus::Paused);
             Core::System::GetInstance().CpuCore(processor_id).PrepareReschedule();
         }
     } else if (status == ThreadStatus::Paused) {
@@ -408,6 +360,170 @@ void Thread::Sleep(s64 nanoseconds) {
     WakeAfterDelay(nanoseconds);
 }
 
+bool Thread::YieldSimple() {
+    auto& scheduler = kernel.GlobalScheduler();
+    return scheduler.YieldThread(this);
+}
+
+bool Thread::YieldAndBalanceLoad() {
+    auto& scheduler = kernel.GlobalScheduler();
+    return scheduler.YieldThreadAndBalanceLoad(this);
+}
+
+bool Thread::YieldAndWaitForLoadBalancing() {
+    auto& scheduler = kernel.GlobalScheduler();
+    return scheduler.YieldThreadAndWaitForLoadBalancing(this);
+}
+
+void Thread::SetSchedulingStatus(ThreadSchedStatus new_status) {
+    const u32 old_flags = scheduling_state;
+    scheduling_state = (scheduling_state & static_cast<u32>(ThreadSchedMasks::HighMask)) |
+                       static_cast<u32>(new_status);
+    AdjustSchedulingOnStatus(old_flags);
+}
+
+void Thread::SetCurrentPriority(u32 new_priority) {
+    const u32 old_priority = std::exchange(current_priority, new_priority);
+    AdjustSchedulingOnPriority(old_priority);
+}
+
+ResultCode Thread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) {
+    const auto HighestSetCore = [](u64 mask, u32 max_cores) {
+        for (s32 core = static_cast<s32>(max_cores - 1); core >= 0; core--) {
+            if (((mask >> core) & 1) != 0) {
+                return core;
+            }
+        }
+        return -1;
+    };
+
+    const bool use_override = affinity_override_count != 0;
+    if (new_core == THREADPROCESSORID_DONT_UPDATE) {
+        new_core = use_override ? ideal_core_override : ideal_core;
+        if ((new_affinity_mask & (1ULL << new_core)) == 0) {
+            return ERR_INVALID_COMBINATION;
+        }
+    }
+    if (use_override) {
+        ideal_core_override = new_core;
+        affinity_mask_override = new_affinity_mask;
+    } else {
+        const u64 old_affinity_mask = std::exchange(affinity_mask, new_affinity_mask);
+        ideal_core = new_core;
+        if (old_affinity_mask != new_affinity_mask) {
+            const s32 old_core = processor_id;
+            if (processor_id >= 0 && ((affinity_mask >> processor_id) & 1) == 0) {
+                if (static_cast<s32>(ideal_core) < 0) {
+                    processor_id = HighestSetCore(affinity_mask, GlobalScheduler::NUM_CPU_CORES);
+                } else {
+                    processor_id = ideal_core;
+                }
+            }
+            AdjustSchedulingOnAffinity(old_affinity_mask, old_core);
+        }
+    }
+    return RESULT_SUCCESS;
+}
+
+void Thread::AdjustSchedulingOnStatus(u32 old_flags) {
+    if (old_flags == scheduling_state) {
+        return;
+    }
+
+    auto& scheduler = kernel.GlobalScheduler();
+    if (static_cast<ThreadSchedStatus>(old_flags & static_cast<u32>(ThreadSchedMasks::LowMask)) ==
+        ThreadSchedStatus::Runnable) {
+        // In this case the thread was running, now it's pausing/exitting
+        if (processor_id >= 0) {
+            scheduler.Unschedule(current_priority, static_cast<u32>(processor_id), this);
+        }
+
+        for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+            if (core != static_cast<u32>(processor_id) && ((affinity_mask >> core) & 1) != 0) {
+                scheduler.Unsuggest(current_priority, core, this);
+            }
+        }
+    } else if (GetSchedulingStatus() == ThreadSchedStatus::Runnable) {
+        // The thread is now set to running from being stopped
+        if (processor_id >= 0) {
+            scheduler.Schedule(current_priority, static_cast<u32>(processor_id), this);
+        }
+
+        for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+            if (core != static_cast<u32>(processor_id) && ((affinity_mask >> core) & 1) != 0) {
+                scheduler.Suggest(current_priority, core, this);
+            }
+        }
+    }
+
+    scheduler.SetReselectionPending();
+}
+
+void Thread::AdjustSchedulingOnPriority(u32 old_priority) {
+    if (GetSchedulingStatus() != ThreadSchedStatus::Runnable) {
+        return;
+    }
+    auto& scheduler = Core::System::GetInstance().GlobalScheduler();
+    if (processor_id >= 0) {
+        scheduler.Unschedule(old_priority, static_cast<u32>(processor_id), this);
+    }
+
+    for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+        if (core != static_cast<u32>(processor_id) && ((affinity_mask >> core) & 1) != 0) {
+            scheduler.Unsuggest(old_priority, core, this);
+        }
+    }
+
+    // Add thread to the new priority queues.
+    Thread* current_thread = GetCurrentThread();
+
+    if (processor_id >= 0) {
+        if (current_thread == this) {
+            scheduler.SchedulePrepend(current_priority, static_cast<u32>(processor_id), this);
+        } else {
+            scheduler.Schedule(current_priority, static_cast<u32>(processor_id), this);
+        }
+    }
+
+    for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+        if (core != static_cast<u32>(processor_id) && ((affinity_mask >> core) & 1) != 0) {
+            scheduler.Suggest(current_priority, core, this);
+        }
+    }
+
+    scheduler.SetReselectionPending();
+}
+
+void Thread::AdjustSchedulingOnAffinity(u64 old_affinity_mask, s32 old_core) {
+    auto& scheduler = Core::System::GetInstance().GlobalScheduler();
+    if (GetSchedulingStatus() != ThreadSchedStatus::Runnable ||
+        current_priority >= THREADPRIO_COUNT) {
+        return;
+    }
+
+    for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+        if (((old_affinity_mask >> core) & 1) != 0) {
+            if (core == static_cast<u32>(old_core)) {
+                scheduler.Unschedule(current_priority, core, this);
+            } else {
+                scheduler.Unsuggest(current_priority, core, this);
+            }
+        }
+    }
+
+    for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+        if (((affinity_mask >> core) & 1) != 0) {
+            if (core == static_cast<u32>(processor_id)) {
+                scheduler.Schedule(current_priority, core, this);
+            } else {
+                scheduler.Suggest(current_priority, core, this);
+            }
+        }
+    }
+
+    scheduler.SetReselectionPending();
+}
+
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 
 /**

src/core/hle/kernel/thread.h

@@ -75,6 +75,26 @@ enum class ThreadActivity : u32 {
     Paused = 1,
 };
 
+enum class ThreadSchedStatus : u32 {
+    None = 0,
+    Paused = 1,
+    Runnable = 2,
+    Exited = 3,
+};
+
+enum class ThreadSchedFlags : u32 {
+    ProcessPauseFlag = 1 << 4,
+    ThreadPauseFlag = 1 << 5,
+    ProcessDebugPauseFlag = 1 << 6,
+    KernelInitPauseFlag = 1 << 8,
+};
+
+enum class ThreadSchedMasks : u32 {
+    LowMask = 0x000f,
+    HighMask = 0xfff0,
+    ForcePauseMask = 0x0070,
+};
+
 class Thread final : public WaitObject {
 public:
     using MutexWaitingThreads = std::vector<SharedPtr<Thread>>;
@@ -278,6 +298,10 @@ public:
         return processor_id;
     }
 
+    void SetProcessorID(s32 new_core) {
+        processor_id = new_core;
+    }
+
     Process* GetOwnerProcess() {
         return owner_process;
     }
@@ -295,6 +319,9 @@ public:
     }
 
     void ClearWaitObjects() {
+        for (const auto& waiting_object : wait_objects) {
+            waiting_object->RemoveWaitingThread(this);
+        }
         wait_objects.clear();
     }
 
@@ -383,11 +410,47 @@ public:
     /// Sleeps this thread for the given amount of nanoseconds.
     void Sleep(s64 nanoseconds);
 
+    /// Yields this thread without rebalancing loads.
+    bool YieldSimple();
+
+    /// Yields this thread and does a load rebalancing.
+    bool YieldAndBalanceLoad();
+
+    /// Yields this thread and if the core is left idle, loads are rebalanced
+    bool YieldAndWaitForLoadBalancing();
+
+    void IncrementYieldCount() {
+        yield_count++;
+    }
+
+    u64 GetYieldCount() const {
+        return yield_count;
+    }
+
+    ThreadSchedStatus GetSchedulingStatus() const {
+        return static_cast<ThreadSchedStatus>(scheduling_state &
+                                              static_cast<u32>(ThreadSchedMasks::LowMask));
+    }
+
+    bool IsRunning() const {
+        return is_running;
+    }
+
+    void SetIsRunning(bool value) {
+        is_running = value;
+    }
+
 private:
     explicit Thread(KernelCore& kernel);
     ~Thread() override;
 
-    void ChangeScheduler();
+    void SetSchedulingStatus(ThreadSchedStatus new_status);
+    void SetCurrentPriority(u32 new_priority);
+    ResultCode SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask);
+
+    void AdjustSchedulingOnStatus(u32 old_flags);
+    void AdjustSchedulingOnPriority(u32 old_priority);
+    void AdjustSchedulingOnAffinity(u64 old_affinity_mask, s32 old_core);
 
     Core::ARM_Interface::ThreadContext context{};
 
@@ -409,6 +472,8 @@ private:
 
     u64 total_cpu_time_ticks = 0; ///< Total CPU running ticks.
     u64 last_running_ticks = 0;   ///< CPU tick when thread was last running
+    u64 yield_count = 0;          ///< Number of redundant yields carried by this thread.
+                                  ///< a redundant yield is one where no scheduling is changed
 
     s32 processor_id = 0;
 
@@ -453,6 +518,13 @@ private:
 
     ThreadActivity activity = ThreadActivity::Normal;
 
+    s32 ideal_core_override = -1;
+    u64 affinity_mask_override = 0x1;
+    u32 affinity_override_count = 0;
+
+    u32 scheduling_state = 0;
+    bool is_running = false;
+
     std::string name;
 };
 

src/core/hle/kernel/vm_manager.cpp

@@ -167,7 +167,7 @@ ResultVal<VAddr> VMManager::FindFreeRegion(VAddr begin, VAddr end, u64 size) con
 
     if (vma_handle == vma_map.cend()) {
         // TODO(Subv): Find the correct error code here.
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
 
     const VAddr target = std::max(begin, vma_handle->second.base);

src/core/hle/kernel/wait_object.cpp

@@ -6,6 +6,9 @@
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "common/logging/log.h"
+#include "core/core.h"
+#include "core/core_cpu.h"
+#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/thread.h"
@@ -82,9 +85,6 @@ void WaitObject::WakeupWaitingThread(SharedPtr<Thread> thread) {
 
     const std::size_t index = thread->GetWaitObjectIndex(this);
 
-    for (const auto& object : thread->GetWaitObjects()) {
-        object->RemoveWaitingThread(thread.get());
-    }
     thread->ClearWaitObjects();
 
     thread->CancelWakeupTimer();
@@ -95,6 +95,7 @@ void WaitObject::WakeupWaitingThread(SharedPtr<Thread> thread) {
     }
     if (resume) {
         thread->ResumeFromWait();
+        Core::System::GetInstance().PrepareReschedule(thread->GetProcessorID());
     }
 }
 

src/core/hle/kernel/writable_event.cpp

@@ -15,8 +15,7 @@ namespace Kernel {
 WritableEvent::WritableEvent(KernelCore& kernel) : Object{kernel} {}
 WritableEvent::~WritableEvent() = default;
 
-EventPair WritableEvent::CreateEventPair(KernelCore& kernel, ResetType reset_type,
-                                         std::string name) {
+EventPair WritableEvent::CreateEventPair(KernelCore& kernel, std::string name) {
     SharedPtr<WritableEvent> writable_event(new WritableEvent(kernel));
     SharedPtr<ReadableEvent> readable_event(new ReadableEvent(kernel));
 
@@ -24,7 +23,6 @@ EventPair WritableEvent::CreateEventPair(KernelCore& kernel, ResetType reset_typ
     writable_event->readable = readable_event;
     readable_event->name = name + ":Readable";
     readable_event->signaled = false;
-    readable_event->reset_type = reset_type;
 
     return {std::move(readable_event), std::move(writable_event)};
 }
@@ -33,10 +31,6 @@ SharedPtr<ReadableEvent> WritableEvent::GetReadableEvent() const {
     return readable;
 }
 
-ResetType WritableEvent::GetResetType() const {
-    return readable->reset_type;
-}
-
 void WritableEvent::Signal() {
     readable->Signal();
 }

src/core/hle/kernel/writable_event.h

@@ -24,11 +24,9 @@ public:
     /**
      * Creates an event
      * @param kernel The kernel instance to create this event under.
-     * @param reset_type ResetType describing how to create event
      * @param name Optional name of event
      */
-    static EventPair CreateEventPair(KernelCore& kernel, ResetType reset_type,
-                                     std::string name = "Unknown");
+    static EventPair CreateEventPair(KernelCore& kernel, std::string name = "Unknown");
 
     std::string GetTypeName() const override {
         return "WritableEvent";
@@ -44,8 +42,6 @@ public:
 
     SharedPtr<ReadableEvent> GetReadableEvent() const;
 
-    ResetType GetResetType() const;
-
     void Signal();
     void Clear();
     bool IsSignaled() const;

src/core/hle/result.h

@@ -146,6 +146,14 @@ constexpr bool operator!=(const ResultCode& a, const ResultCode& b) {
 /// The default success `ResultCode`.
 constexpr ResultCode RESULT_SUCCESS(0);
 
+/**
+ * Placeholder result code used for unknown error codes.
+ *
+ * @note This should only be used when a particular error code
+ *       is not known yet.
+ */
+constexpr ResultCode RESULT_UNKNOWN(UINT32_MAX);
+
 /**
  * This is an optional value type. It holds a `ResultCode` and, if that code is a success code,
  * also holds a result of type `T`. If the code is an error code then trying to access the inner
@@ -183,7 +191,7 @@ class ResultVal {
 public:
     /// Constructs an empty `ResultVal` with the given error code. The code must not be a success
     /// code.
-    ResultVal(ResultCode error_code = ResultCode(-1)) : result_code(error_code) {
+    ResultVal(ResultCode error_code = RESULT_UNKNOWN) : result_code(error_code) {
         ASSERT(error_code.IsError());
     }
 

src/core/hle/service/acc/acc.cpp

@@ -84,7 +84,7 @@ protected:
             LOG_ERROR(Service_ACC, "Failed to get profile base and data for user={}",
                       user_id.Format());
             IPC::ResponseBuilder rb{ctx, 2};
-            rb.Push(ResultCode(-1)); // TODO(ogniK): Get actual error code
+            rb.Push(RESULT_UNKNOWN); // TODO(ogniK): Get actual error code
         }
     }
 
@@ -98,7 +98,7 @@ protected:
         } else {
             LOG_ERROR(Service_ACC, "Failed to get profile base for user={}", user_id.Format());
             IPC::ResponseBuilder rb{ctx, 2};
-            rb.Push(ResultCode(-1)); // TODO(ogniK): Get actual error code
+            rb.Push(RESULT_UNKNOWN); // TODO(ogniK): Get actual error code
         }
     }
 
@@ -442,7 +442,7 @@ void Module::Interface::TrySelectUserWithoutInteraction(Kernel::HLERequestContex
     const auto user_list = profile_manager->GetAllUsers();
     if (std::all_of(user_list.begin(), user_list.end(),
                     [](const auto& user) { return user.uuid == Common::INVALID_UUID; })) {
-        rb.Push(ResultCode(-1)); // TODO(ogniK): Find the correct error code
+        rb.Push(RESULT_UNKNOWN); // TODO(ogniK): Find the correct error code
         rb.PushRaw<u128>(Common::INVALID_UUID);
         return;
     }

src/core/hle/service/acc/profile_manager.cpp

@@ -31,8 +31,8 @@ struct ProfileDataRaw {
 static_assert(sizeof(ProfileDataRaw) == 0x650, "ProfileDataRaw has incorrect size.");
 
 // TODO(ogniK): Get actual error codes
-constexpr ResultCode ERROR_TOO_MANY_USERS(ErrorModule::Account, -1);
-constexpr ResultCode ERROR_USER_ALREADY_EXISTS(ErrorModule::Account, -2);
+constexpr ResultCode ERROR_TOO_MANY_USERS(ErrorModule::Account, u32(-1));
+constexpr ResultCode ERROR_USER_ALREADY_EXISTS(ErrorModule::Account, u32(-2));
 constexpr ResultCode ERROR_ARGUMENT_IS_NULL(ErrorModule::Account, 20);
 
 constexpr char ACC_SAVE_AVATORS_BASE_PATH[] = "/system/save/8000000000000010/su/avators/";

src/core/hle/service/am/am.cpp

@@ -289,8 +289,8 @@ ISelfController::ISelfController(Core::System& system,
     RegisterHandlers(functions);
 
     auto& kernel = system.Kernel();
-    launchable_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                              "ISelfController:LaunchableEvent");
+    launchable_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, "ISelfController:LaunchableEvent");
 
     // This event is created by AM on the first time GetAccumulatedSuspendedTickChangedEvent() is
     // called. Yuzu can just create it unconditionally, since it doesn't need to support multiple
@@ -298,7 +298,7 @@ ISelfController::ISelfController(Core::System& system,
     // suspended if the event has previously been created by a call to
     // GetAccumulatedSuspendedTickChangedEvent.
     accumulated_suspended_tick_changed_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "ISelfController:AccumulatedSuspendedTickChangedEvent");
+        kernel, "ISelfController:AccumulatedSuspendedTickChangedEvent");
     accumulated_suspended_tick_changed_event.writable->Signal();
 }
 
@@ -523,10 +523,10 @@ void ISelfController::GetAccumulatedSuspendedTickChangedEvent(Kernel::HLERequest
 }
 
 AppletMessageQueue::AppletMessageQueue(Kernel::KernelCore& kernel) {
-    on_new_message = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                            "AMMessageQueue:OnMessageRecieved");
-    on_operation_mode_changed = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Automatic, "AMMessageQueue:OperationModeChanged");
+    on_new_message =
+        Kernel::WritableEvent::CreateEventPair(kernel, "AMMessageQueue:OnMessageRecieved");
+    on_operation_mode_changed =
+        Kernel::WritableEvent::CreateEventPair(kernel, "AMMessageQueue:OperationModeChanged");
 }
 
 AppletMessageQueue::~AppletMessageQueue() = default;
@@ -991,7 +991,7 @@ void ILibraryAppletCreator::CreateLibraryApplet(Kernel::HLERequestContext& ctx)
         LOG_ERROR(Service_AM, "Applet doesn't exist! applet_id={}", static_cast<u32>(applet_id));
 
         IPC::ResponseBuilder rb{ctx, 2};
-        rb.Push(ResultCode(-1));
+        rb.Push(RESULT_UNKNOWN);
         return;
     }
 
@@ -1027,7 +1027,7 @@ void ILibraryAppletCreator::CreateTransferMemoryStorage(Kernel::HLERequestContex
     if (transfer_mem == nullptr) {
         LOG_ERROR(Service_AM, "shared_mem is a nullpr for handle={:08X}", handle);
         IPC::ResponseBuilder rb{ctx, 2};
-        rb.Push(ResultCode(-1));
+        rb.Push(RESULT_UNKNOWN);
         return;
     }
 
@@ -1073,9 +1073,9 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_)
         {71, nullptr, "RequestToReboot"},
         {80, nullptr, "ExitAndRequestToShowThanksMessage"},
         {90, &IApplicationFunctions::EnableApplicationCrashReport, "EnableApplicationCrashReport"},
-        {100, nullptr, "InitializeApplicationCopyrightFrameBuffer"},
-        {101, nullptr, "SetApplicationCopyrightImage"},
-        {102, nullptr, "SetApplicationCopyrightVisibility"},
+        {100, &IApplicationFunctions::InitializeApplicationCopyrightFrameBuffer, "InitializeApplicationCopyrightFrameBuffer"},
+        {101, &IApplicationFunctions::SetApplicationCopyrightImage, "SetApplicationCopyrightImage"},
+        {102, &IApplicationFunctions::SetApplicationCopyrightVisibility, "SetApplicationCopyrightVisibility"},
         {110, nullptr, "QueryApplicationPlayStatistics"},
         {120, nullptr, "ExecuteProgram"},
         {121, nullptr, "ClearUserChannel"},
@@ -1091,7 +1091,7 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_)
 
     auto& kernel = system.Kernel();
     gpu_error_detected_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "IApplicationFunctions:GpuErrorDetectedSystemEvent");
+        kernel, "IApplicationFunctions:GpuErrorDetectedSystemEvent");
 }
 
 IApplicationFunctions::~IApplicationFunctions() = default;
@@ -1103,6 +1103,31 @@ void IApplicationFunctions::EnableApplicationCrashReport(Kernel::HLERequestConte
     rb.Push(RESULT_SUCCESS);
 }
 
+void IApplicationFunctions::InitializeApplicationCopyrightFrameBuffer(
+    Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void IApplicationFunctions::SetApplicationCopyrightImage(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void IApplicationFunctions::SetApplicationCopyrightVisibility(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto is_visible = rp.Pop<bool>();
+
+    LOG_WARNING(Service_AM, "(STUBBED) called, is_visible={}", is_visible);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
 void IApplicationFunctions::BeginBlockingHomeButtonShortAndLongPressed(
     Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_AM, "(STUBBED) called");

src/core/hle/service/am/am.h

@@ -252,6 +252,9 @@ private:
     void BeginBlockingHomeButton(Kernel::HLERequestContext& ctx);
     void EndBlockingHomeButton(Kernel::HLERequestContext& ctx);
     void EnableApplicationCrashReport(Kernel::HLERequestContext& ctx);
+    void InitializeApplicationCopyrightFrameBuffer(Kernel::HLERequestContext& ctx);
+    void SetApplicationCopyrightImage(Kernel::HLERequestContext& ctx);
+    void SetApplicationCopyrightVisibility(Kernel::HLERequestContext& ctx);
     void GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx);
 
     bool launch_popped_application_specific = false;

src/core/hle/service/am/applets/applets.cpp

@@ -24,12 +24,12 @@
 namespace Service::AM::Applets {
 
 AppletDataBroker::AppletDataBroker(Kernel::KernelCore& kernel) {
-    state_changed_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "ILibraryAppletAccessor:StateChangedEvent");
-    pop_out_data_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "ILibraryAppletAccessor:PopDataOutEvent");
+    state_changed_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, "ILibraryAppletAccessor:StateChangedEvent");
+    pop_out_data_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, "ILibraryAppletAccessor:PopDataOutEvent");
     pop_interactive_out_data_event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Manual, "ILibraryAppletAccessor:PopInteractiveDataOutEvent");
+        kernel, "ILibraryAppletAccessor:PopInteractiveDataOutEvent");
 }
 
 AppletDataBroker::~AppletDataBroker() = default;

src/core/hle/service/am/applets/error.cpp

@@ -20,9 +20,9 @@ namespace Service::AM::Applets {
 struct ShowError {
     u8 mode;
     bool jump;
-    INSERT_PADDING_BYTES(4);
+    INSERT_UNION_PADDING_BYTES(4);
     bool use_64bit_error_code;
-    INSERT_PADDING_BYTES(1);
+    INSERT_UNION_PADDING_BYTES(1);
     u64 error_code_64;
     u32 error_code_32;
 };
@@ -32,7 +32,7 @@ static_assert(sizeof(ShowError) == 0x14, "ShowError has incorrect size.");
 struct ShowErrorRecord {
     u8 mode;
     bool jump;
-    INSERT_PADDING_BYTES(6);
+    INSERT_UNION_PADDING_BYTES(6);
     u64 error_code_64;
     u64 posix_time;
 };
@@ -41,7 +41,7 @@ static_assert(sizeof(ShowErrorRecord) == 0x18, "ShowErrorRecord has incorrect si
 struct SystemErrorArg {
     u8 mode;
     bool jump;
-    INSERT_PADDING_BYTES(6);
+    INSERT_UNION_PADDING_BYTES(6);
     u64 error_code_64;
     std::array<char, 8> language_code;
     std::array<char, 0x800> main_text;
@@ -52,7 +52,7 @@ static_assert(sizeof(SystemErrorArg) == 0x1018, "SystemErrorArg has incorrect si
 struct ApplicationErrorArg {
     u8 mode;
     bool jump;
-    INSERT_PADDING_BYTES(6);
+    INSERT_UNION_PADDING_BYTES(6);
     u32 error_code;
     std::array<char, 8> language_code;
     std::array<char, 0x800> main_text;
@@ -65,6 +65,7 @@ union Error::ErrorArguments {
     ShowErrorRecord error_record;
     SystemErrorArg system_error;
     ApplicationErrorArg application_error;
+    std::array<u8, 0x1018> raw{};
 };
 
 namespace {

src/core/hle/service/am/applets/web_browser.cpp

@@ -337,7 +337,7 @@ void WebBrowser::ExecuteInternal() {
 void WebBrowser::InitializeShop() {
     if (frontend_e_commerce == nullptr) {
         LOG_ERROR(Service_AM, "Missing ECommerce Applet frontend!");
-        status = ResultCode(-1);
+        status = RESULT_UNKNOWN;
         return;
     }
 
@@ -353,7 +353,7 @@ void WebBrowser::InitializeShop() {
 
     if (url == args.end()) {
         LOG_ERROR(Service_AM, "Missing EShop Arguments URL for initialization!");
-        status = ResultCode(-1);
+        status = RESULT_UNKNOWN;
         return;
     }
 
@@ -366,7 +366,7 @@ void WebBrowser::InitializeShop() {
     // Less is missing info, More is malformed
     if (split_query.size() != 2) {
         LOG_ERROR(Service_AM, "EShop Arguments has more than one question mark, malformed");
-        status = ResultCode(-1);
+        status = RESULT_UNKNOWN;
         return;
     }
 
@@ -390,7 +390,7 @@ void WebBrowser::InitializeShop() {
 
     if (scene == shop_query.end()) {
         LOG_ERROR(Service_AM, "No scene parameter was passed via shop query!");
-        status = ResultCode(-1);
+        status = RESULT_UNKNOWN;
         return;
     }
 
@@ -406,7 +406,7 @@ void WebBrowser::InitializeShop() {
     const auto target = target_map.find(scene->second);
     if (target == target_map.end()) {
         LOG_ERROR(Service_AM, "Scene for shop query is invalid! (scene={})", scene->second);
-        status = ResultCode(-1);
+        status = RESULT_UNKNOWN;
         return;
     }
 
@@ -427,7 +427,7 @@ void WebBrowser::InitializeOffline() {
     if (args.find(WebArgTLVType::DocumentPath) == args.end() ||
         args.find(WebArgTLVType::DocumentKind) == args.end() ||
         args.find(WebArgTLVType::ApplicationID) == args.end()) {
-        status = ResultCode(-1);
+        status = RESULT_UNKNOWN;
         LOG_ERROR(Service_AM, "Missing necessary parameters for initialization!");
     }
 
@@ -476,7 +476,7 @@ void WebBrowser::InitializeOffline() {
 
     offline_romfs = GetApplicationRomFS(system, title_id, type);
     if (offline_romfs == nullptr) {
-        status = ResultCode(-1);
+        status = RESULT_UNKNOWN;
         LOG_ERROR(Service_AM, "Failed to find offline data for request!");
     }
 
@@ -496,7 +496,7 @@ void WebBrowser::ExecuteShop() {
     const auto check_optional_parameter = [this](const auto& p) {
         if (!p.has_value()) {
             LOG_ERROR(Service_AM, "Missing one or more necessary parameters for execution!");
-            status = ResultCode(-1);
+            status = RESULT_UNKNOWN;
             return false;
         }
 

src/core/hle/service/aoc/aoc_u.cpp

@@ -67,8 +67,8 @@ AOC_U::AOC_U(Core::System& system)
     RegisterHandlers(functions);
 
     auto& kernel = system.Kernel();
-    aoc_change_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                              "GetAddOnContentListChanged:Event");
+    aoc_change_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, "GetAddOnContentListChanged:Event");
 }
 
 AOC_U::~AOC_U() = default;
@@ -131,7 +131,7 @@ void AOC_U::ListAddOnContent(Kernel::HLERequestContext& ctx) {
     if (out.size() < offset) {
         IPC::ResponseBuilder rb{ctx, 2};
         // TODO(DarkLordZach): Find the correct error code.
-        rb.Push(ResultCode(-1));
+        rb.Push(RESULT_UNKNOWN);
         return;
     }
 

src/core/hle/service/audio/audout_u.cpp

@@ -65,8 +65,8 @@ public:
         RegisterHandlers(functions);
 
         // This is the event handle used to check if the audio buffer was released
-        buffer_event = Kernel::WritableEvent::CreateEventPair(
-            system.Kernel(), Kernel::ResetType::Manual, "IAudioOutBufferReleased");
+        buffer_event =
+            Kernel::WritableEvent::CreateEventPair(system.Kernel(), "IAudioOutBufferReleased");
 
         stream = audio_core.OpenStream(system.CoreTiming(), audio_params.sample_rate,
                                        audio_params.channel_count, std::move(unique_name),

src/core/hle/service/audio/audren_u.cpp

@@ -47,8 +47,8 @@ public:
         // clang-format on
         RegisterHandlers(functions);
 
-        system_event = Kernel::WritableEvent::CreateEventPair(
-            system.Kernel(), Kernel::ResetType::Manual, "IAudioRenderer:SystemEvent");
+        system_event =
+            Kernel::WritableEvent::CreateEventPair(system.Kernel(), "IAudioRenderer:SystemEvent");
         renderer = std::make_unique<AudioCore::AudioRenderer>(
             system.CoreTiming(), audren_params, system_event.writable, instance_number);
     }
@@ -180,17 +180,17 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        buffer_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                              "IAudioOutBufferReleasedEvent");
+        buffer_event =
+            Kernel::WritableEvent::CreateEventPair(kernel, "IAudioOutBufferReleasedEvent");
 
         // Should be similar to audio_output_device_switch_event
         audio_input_device_switch_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IAudioDevice:AudioInputDeviceSwitchedEvent");
+            kernel, "IAudioDevice:AudioInputDeviceSwitchedEvent");
 
         // Should only be signalled when an audio output device has been changed, example: speaker
         // to headset
         audio_output_device_switch_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IAudioDevice:AudioOutputDeviceSwitchedEvent");
+            kernel, "IAudioDevice:AudioOutputDeviceSwitchedEvent");
     }
 
 private:

src/core/hle/service/audio/hwopus.cpp

@@ -80,7 +80,7 @@ private:
             LOG_ERROR(Audio, "Failed to decode opus data");
             IPC::ResponseBuilder rb{ctx, 2};
             // TODO(ogniK): Use correct error code
-            rb.Push(ResultCode(-1));
+            rb.Push(RESULT_UNKNOWN);
             return;
         }
 
@@ -278,7 +278,7 @@ void HwOpus::OpenOpusDecoder(Kernel::HLERequestContext& ctx) {
         LOG_ERROR(Audio, "Failed to create Opus decoder (error={}).", error);
         IPC::ResponseBuilder rb{ctx, 2};
         // TODO(ogniK): Use correct error code
-        rb.Push(ResultCode(-1));
+        rb.Push(RESULT_UNKNOWN);
         return;
     }
 

src/core/hle/service/bcat/backend/backend.cpp

@@ -13,8 +13,7 @@ namespace Service::BCAT {
 ProgressServiceBackend::ProgressServiceBackend(Kernel::KernelCore& kernel,
                                                std::string_view event_name) {
     event = Kernel::WritableEvent::CreateEventPair(
-        kernel, Kernel::ResetType::Automatic,
-        std::string("ProgressServiceBackend:UpdateEvent:").append(event_name));
+        kernel, std::string("ProgressServiceBackend:UpdateEvent:").append(event_name));
 }
 
 Kernel::SharedPtr<Kernel::ReadableEvent> ProgressServiceBackend::GetEvent() const {

src/core/hle/service/bcat/backend/boxcat.cpp

@@ -114,7 +114,7 @@ void HandleDownloadDisplayResult(const AM::Applets::AppletManager& applet_manage
 
     const auto& frontend{applet_manager.GetAppletFrontendSet()};
     frontend.error->ShowCustomErrorText(
-        ResultCode(-1), "There was an error while attempting to use Boxcat.",
+        RESULT_UNKNOWN, "There was an error while attempting to use Boxcat.",
         DOWNLOAD_RESULT_LOG_MESSAGES[static_cast<std::size_t>(res)], [] {});
 }
 

src/core/hle/service/btdrv/btdrv.cpp

@@ -34,8 +34,7 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        register_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "BT:RegisterEvent");
+        register_event = Kernel::WritableEvent::CreateEventPair(kernel, "BT:RegisterEvent");
     }
 
 private:

src/core/hle/service/btm/btm.cpp

@@ -57,14 +57,12 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        scan_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                            "IBtmUserCore:ScanEvent");
-        connection_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IBtmUserCore:ConnectionEvent");
-        service_discovery = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IBtmUserCore:Discovery");
-        config_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                              "IBtmUserCore:ConfigEvent");
+        scan_event = Kernel::WritableEvent::CreateEventPair(kernel, "IBtmUserCore:ScanEvent");
+        connection_event =
+            Kernel::WritableEvent::CreateEventPair(kernel, "IBtmUserCore:ConnectionEvent");
+        service_discovery =
+            Kernel::WritableEvent::CreateEventPair(kernel, "IBtmUserCore:Discovery");
+        config_event = Kernel::WritableEvent::CreateEventPair(kernel, "IBtmUserCore:ConfigEvent");
     }
 
 private:

src/core/hle/service/filesystem/filesystem.cpp

@@ -58,11 +58,11 @@ ResultCode VfsDirectoryServiceWrapper::CreateFile(const std::string& path_, u64
     auto file = dir->CreateFile(FileUtil::GetFilename(path));
     if (file == nullptr) {
         // TODO(DarkLordZach): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
     if (!file->Resize(size)) {
         // TODO(DarkLordZach): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
     return RESULT_SUCCESS;
 }
@@ -80,7 +80,7 @@ ResultCode VfsDirectoryServiceWrapper::DeleteFile(const std::string& path_) cons
     }
     if (!dir->DeleteFile(FileUtil::GetFilename(path))) {
         // TODO(DarkLordZach): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
 
     return RESULT_SUCCESS;
@@ -94,7 +94,7 @@ ResultCode VfsDirectoryServiceWrapper::CreateDirectory(const std::string& path_)
     auto new_dir = dir->CreateSubdirectory(FileUtil::GetFilename(path));
     if (new_dir == nullptr) {
         // TODO(DarkLordZach): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
     return RESULT_SUCCESS;
 }
@@ -104,7 +104,7 @@ ResultCode VfsDirectoryServiceWrapper::DeleteDirectory(const std::string& path_)
     auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(path));
     if (!dir->DeleteSubdirectory(FileUtil::GetFilename(path))) {
         // TODO(DarkLordZach): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
     return RESULT_SUCCESS;
 }
@@ -114,7 +114,7 @@ ResultCode VfsDirectoryServiceWrapper::DeleteDirectoryRecursively(const std::str
     auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(path));
     if (!dir->DeleteSubdirectoryRecursive(FileUtil::GetFilename(path))) {
         // TODO(DarkLordZach): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
     return RESULT_SUCCESS;
 }
@@ -125,7 +125,7 @@ ResultCode VfsDirectoryServiceWrapper::CleanDirectoryRecursively(const std::stri
 
     if (!dir->CleanSubdirectoryRecursive(FileUtil::GetFilename(sanitized_path))) {
         // TODO(DarkLordZach): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
 
     return RESULT_SUCCESS;
@@ -142,7 +142,7 @@ ResultCode VfsDirectoryServiceWrapper::RenameFile(const std::string& src_path_,
             return FileSys::ERROR_PATH_NOT_FOUND;
         if (!src->Rename(FileUtil::GetFilename(dest_path))) {
             // TODO(DarkLordZach): Find a better error code for this
-            return ResultCode(-1);
+            return RESULT_UNKNOWN;
         }
         return RESULT_SUCCESS;
     }
@@ -160,7 +160,7 @@ ResultCode VfsDirectoryServiceWrapper::RenameFile(const std::string& src_path_,
 
     if (!src->GetContainingDirectory()->DeleteFile(FileUtil::GetFilename(src_path))) {
         // TODO(DarkLordZach): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
 
     return RESULT_SUCCESS;
@@ -177,7 +177,7 @@ ResultCode VfsDirectoryServiceWrapper::RenameDirectory(const std::string& src_pa
             return FileSys::ERROR_PATH_NOT_FOUND;
         if (!src->Rename(FileUtil::GetFilename(dest_path))) {
             // TODO(DarkLordZach): Find a better error code for this
-            return ResultCode(-1);
+            return RESULT_UNKNOWN;
         }
         return RESULT_SUCCESS;
     }
@@ -189,7 +189,7 @@ ResultCode VfsDirectoryServiceWrapper::RenameDirectory(const std::string& src_pa
                src_path, dest_path);
 
     // TODO(DarkLordZach): Find a better error code for this
-    return ResultCode(-1);
+    return RESULT_UNKNOWN;
 }
 
 ResultVal<FileSys::VirtualFile> VfsDirectoryServiceWrapper::OpenFile(const std::string& path_,
@@ -287,7 +287,7 @@ ResultVal<FileSys::VirtualFile> FileSystemController::OpenRomFSCurrentProcess()
 
     if (romfs_factory == nullptr) {
         // TODO(bunnei): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
 
     return romfs_factory->OpenCurrentProcess(system.CurrentProcess()->GetTitleID());
@@ -300,7 +300,7 @@ ResultVal<FileSys::VirtualFile> FileSystemController::OpenRomFS(
 
     if (romfs_factory == nullptr) {
         // TODO(bunnei): Find a better error code for this
-        return ResultCode(-1);
+        return RESULT_UNKNOWN;
     }
 
     return romfs_factory->Open(title_id, storage_id, type);

src/core/hle/service/filesystem/fsp_srv.cpp

@@ -785,7 +785,7 @@ void FSP_SRV::OpenFileSystemWithPatch(Kernel::HLERequestContext& ctx) {
                 static_cast<u8>(type), title_id);
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 0};
-    rb.Push(ResultCode(-1));
+    rb.Push(RESULT_UNKNOWN);
 }
 
 void FSP_SRV::OpenSdCardFileSystem(Kernel::HLERequestContext& ctx) {
@@ -891,7 +891,7 @@ void FSP_SRV::OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx) {
         // TODO (bunnei): Find the right error code to use here
         LOG_CRITICAL(Service_FS, "no file system interface available!");
         IPC::ResponseBuilder rb{ctx, 2};
-        rb.Push(ResultCode(-1));
+        rb.Push(RESULT_UNKNOWN);
         return;
     }
 
@@ -928,7 +928,7 @@ void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) {
                   "could not open data storage with title_id={:016X}, storage_id={:02X}", title_id,
                   static_cast<u8>(storage_id));
         IPC::ResponseBuilder rb{ctx, 2};
-        rb.Push(ResultCode(-1));
+        rb.Push(RESULT_UNKNOWN);
         return;
     }
 

src/core/hle/service/friend/friend.cpp

@@ -162,7 +162,7 @@ public:
         RegisterHandlers(functions);
 
         notification_event = Kernel::WritableEvent::CreateEventPair(
-            system.Kernel(), Kernel::ResetType::Manual, "INotificationService:NotifyEvent");
+            system.Kernel(), "INotificationService:NotifyEvent");
     }
 
 private:

src/core/hle/service/hid/controllers/npad.cpp

@@ -174,7 +174,7 @@ void Controller_NPad::OnInit() {
     auto& kernel = system.Kernel();
     for (std::size_t i = 0; i < styleset_changed_events.size(); i++) {
         styleset_changed_events[i] = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Manual, fmt::format("npad:NpadStyleSetChanged_{}", i));
+            kernel, fmt::format("npad:NpadStyleSetChanged_{}", i));
     }
 
     if (!IsControllerActivated()) {

src/core/hle/service/hid/hid.cpp

@@ -203,13 +203,13 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) {
         {120, &Hid::SetNpadJoyHoldType, "SetNpadJoyHoldType"},
         {121, &Hid::GetNpadJoyHoldType, "GetNpadJoyHoldType"},
         {122, &Hid::SetNpadJoyAssignmentModeSingleByDefault, "SetNpadJoyAssignmentModeSingleByDefault"},
-        {123, nullptr, "SetNpadJoyAssignmentModeSingleByDefault"},
+        {123, &Hid::SetNpadJoyAssignmentModeSingle, "SetNpadJoyAssignmentModeSingle"},
         {124, &Hid::SetNpadJoyAssignmentModeDual, "SetNpadJoyAssignmentModeDual"},
         {125, &Hid::MergeSingleJoyAsDualJoy, "MergeSingleJoyAsDualJoy"},
         {126, &Hid::StartLrAssignmentMode, "StartLrAssignmentMode"},
         {127, &Hid::StopLrAssignmentMode, "StopLrAssignmentMode"},
         {128, &Hid::SetNpadHandheldActivationMode, "SetNpadHandheldActivationMode"},
-        {129, nullptr, "GetNpadHandheldActivationMode"},
+        {129, &Hid::GetNpadHandheldActivationMode, "GetNpadHandheldActivationMode"},
         {130, &Hid::SwapNpadAssignment, "SwapNpadAssignment"},
         {131, nullptr, "IsUnintendedHomeButtonInputProtectionEnabled"},
         {132, nullptr, "EnableUnintendedHomeButtonInputProtection"},
@@ -557,10 +557,126 @@ void Hid::SetNpadJoyAssignmentModeSingleByDefault(Kernel::HLERequestContext& ctx
     LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", npad_id,
                 applet_resource_user_id);
 
+    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+    controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Single);
+
     IPC::ResponseBuilder rb{ctx, 2};
     rb.Push(RESULT_SUCCESS);
 }
 
+void Hid::SetNpadJoyAssignmentModeSingle(Kernel::HLERequestContext& ctx) {
+    // TODO: Check the differences between this and SetNpadJoyAssignmentModeSingleByDefault
+    IPC::RequestParser rp{ctx};
+    const auto npad_id{rp.Pop<u32>()};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+    const auto npad_joy_device_type{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_HID,
+                "(STUBBED) called, npad_id={}, applet_resource_user_id={}, npad_joy_device_type={}",
+                npad_id, applet_resource_user_id, npad_joy_device_type);
+
+    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+    controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Single);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::SetNpadJoyAssignmentModeDual(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto npad_id{rp.Pop<u32>()};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}", npad_id,
+              applet_resource_user_id);
+
+    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+    controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Dual);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::MergeSingleJoyAsDualJoy(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto unknown_1{rp.Pop<u32>()};
+    const auto unknown_2{rp.Pop<u32>()};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_HID,
+                "(STUBBED) called, unknown_1={}, unknown_2={}, applet_resource_user_id={}",
+                unknown_1, unknown_2, applet_resource_user_id);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::StartLrAssignmentMode(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id);
+    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+    controller.StartLRAssignmentMode();
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::StopLrAssignmentMode(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id);
+    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+    controller.StopLRAssignmentMode();
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+    const auto mode{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}, mode={}",
+                applet_resource_user_id, mode);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::GetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}",
+                applet_resource_user_id);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::SwapNpadAssignment(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto npad_1{rp.Pop<u32>()};
+    const auto npad_2{rp.Pop<u32>()};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, npad_1={}, npad_2={}",
+              applet_resource_user_id, npad_1, npad_2);
+
+    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+    IPC::ResponseBuilder rb{ctx, 2};
+    if (controller.SwapNpadAssignment(npad_1, npad_2)) {
+        rb.Push(RESULT_SUCCESS);
+    } else {
+        LOG_ERROR(Service_HID, "Npads are not connected!");
+        rb.Push(ERR_NPAD_NOT_CONNECTED);
+    }
+}
+
 void Hid::BeginPermitVibrationSession(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto applet_resource_user_id{rp.Pop<u64>()};
@@ -635,47 +751,6 @@ void Hid::GetActualVibrationValue(Kernel::HLERequestContext& ctx) {
         applet_resource->GetController<Controller_NPad>(HidController::NPad).GetLastVibration());
 }
 
-void Hid::SetNpadJoyAssignmentModeDual(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp{ctx};
-    const auto npad_id{rp.Pop<u32>()};
-    const auto applet_resource_user_id{rp.Pop<u64>()};
-
-    LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}", npad_id,
-              applet_resource_user_id);
-
-    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
-    controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Dual);
-
-    IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(RESULT_SUCCESS);
-}
-
-void Hid::MergeSingleJoyAsDualJoy(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp{ctx};
-    const auto unknown_1{rp.Pop<u32>()};
-    const auto unknown_2{rp.Pop<u32>()};
-    const auto applet_resource_user_id{rp.Pop<u64>()};
-
-    LOG_WARNING(Service_HID,
-                "(STUBBED) called, unknown_1={}, unknown_2={}, applet_resource_user_id={}",
-                unknown_1, unknown_2, applet_resource_user_id);
-
-    IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(RESULT_SUCCESS);
-}
-
-void Hid::SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp{ctx};
-    const auto applet_resource_user_id{rp.Pop<u64>()};
-    const auto mode{rp.Pop<u64>()};
-
-    LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}, mode={}",
-                applet_resource_user_id, mode);
-
-    IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(RESULT_SUCCESS);
-}
-
 void Hid::GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_HID, "called");
 
@@ -769,49 +844,6 @@ void Hid::SetPalmaBoostMode(Kernel::HLERequestContext& ctx) {
     rb.Push(RESULT_SUCCESS);
 }
 
-void Hid::StartLrAssignmentMode(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp{ctx};
-    const auto applet_resource_user_id{rp.Pop<u64>()};
-
-    LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id);
-    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
-    controller.StartLRAssignmentMode();
-
-    IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(RESULT_SUCCESS);
-}
-
-void Hid::StopLrAssignmentMode(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp{ctx};
-    const auto applet_resource_user_id{rp.Pop<u64>()};
-
-    LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id);
-    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
-    controller.StopLRAssignmentMode();
-
-    IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(RESULT_SUCCESS);
-}
-
-void Hid::SwapNpadAssignment(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp{ctx};
-    const auto npad_1{rp.Pop<u32>()};
-    const auto npad_2{rp.Pop<u32>()};
-    const auto applet_resource_user_id{rp.Pop<u64>()};
-
-    LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, npad_1={}, npad_2={}",
-              applet_resource_user_id, npad_1, npad_2);
-
-    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
-    IPC::ResponseBuilder rb{ctx, 2};
-    if (controller.SwapNpadAssignment(npad_1, npad_2)) {
-        rb.Push(RESULT_SUCCESS);
-    } else {
-        LOG_ERROR(Service_HID, "Npads are not connected!");
-        rb.Push(ERR_NPAD_NOT_CONNECTED);
-    }
-}
-
 class HidDbg final : public ServiceFramework<HidDbg> {
 public:
     explicit HidDbg() : ServiceFramework{"hid:dbg"} {

src/core/hle/service/hid/hid.h

@@ -106,14 +106,19 @@ private:
     void SetNpadJoyHoldType(Kernel::HLERequestContext& ctx);
     void GetNpadJoyHoldType(Kernel::HLERequestContext& ctx);
     void SetNpadJoyAssignmentModeSingleByDefault(Kernel::HLERequestContext& ctx);
+    void SetNpadJoyAssignmentModeSingle(Kernel::HLERequestContext& ctx);
+    void SetNpadJoyAssignmentModeDual(Kernel::HLERequestContext& ctx);
+    void MergeSingleJoyAsDualJoy(Kernel::HLERequestContext& ctx);
+    void StartLrAssignmentMode(Kernel::HLERequestContext& ctx);
+    void StopLrAssignmentMode(Kernel::HLERequestContext& ctx);
+    void SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx);
+    void GetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx);
+    void SwapNpadAssignment(Kernel::HLERequestContext& ctx);
     void BeginPermitVibrationSession(Kernel::HLERequestContext& ctx);
     void EndPermitVibrationSession(Kernel::HLERequestContext& ctx);
     void SendVibrationValue(Kernel::HLERequestContext& ctx);
     void SendVibrationValues(Kernel::HLERequestContext& ctx);
     void GetActualVibrationValue(Kernel::HLERequestContext& ctx);
-    void SetNpadJoyAssignmentModeDual(Kernel::HLERequestContext& ctx);
-    void MergeSingleJoyAsDualJoy(Kernel::HLERequestContext& ctx);
-    void SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx);
     void GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx);
     void CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx);
     void PermitVibration(Kernel::HLERequestContext& ctx);
@@ -123,9 +128,6 @@ private:
     void StopSixAxisSensor(Kernel::HLERequestContext& ctx);
     void SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx);
     void SetPalmaBoostMode(Kernel::HLERequestContext& ctx);
-    void StartLrAssignmentMode(Kernel::HLERequestContext& ctx);
-    void StopLrAssignmentMode(Kernel::HLERequestContext& ctx);
-    void SwapNpadAssignment(Kernel::HLERequestContext& ctx);
 
     std::shared_ptr<IAppletResource> applet_resource;
     Core::System& system;

src/core/hle/service/mii/mii.cpp

@@ -242,7 +242,7 @@ private:
         const auto index = db.IndexOf(uuid);
         if (index > MAX_MIIS) {
             // TODO(DarkLordZach): Find a better error code
-            rb.Push(ResultCode(-1));
+            rb.Push(RESULT_UNKNOWN);
             rb.Push(index);
         } else {
             rb.Push(RESULT_SUCCESS);
@@ -268,7 +268,7 @@ private:
 
         IPC::ResponseBuilder rb{ctx, 2};
         // TODO(DarkLordZach): Find a better error code
-        rb.Push(success ? RESULT_SUCCESS : ResultCode(-1));
+        rb.Push(success ? RESULT_SUCCESS : RESULT_UNKNOWN);
     }
 
     void AddOrReplace(Kernel::HLERequestContext& ctx) {
@@ -282,7 +282,7 @@ private:
 
         IPC::ResponseBuilder rb{ctx, 2};
         // TODO(DarkLordZach): Find a better error code
-        rb.Push(success ? RESULT_SUCCESS : ResultCode(-1));
+        rb.Push(success ? RESULT_SUCCESS : RESULT_UNKNOWN);
     }
 
     void Delete(Kernel::HLERequestContext& ctx) {

src/core/hle/service/nfp/nfp.cpp

@@ -16,18 +16,14 @@
 #include "core/hle/service/nfp/nfp_user.h"
 
 namespace Service::NFP {
-
 namespace ErrCodes {
-[[maybe_unused]] constexpr ResultCode ERR_TAG_FAILED(ErrorModule::NFP,
-                                                     -1); // TODO(ogniK): Find the actual error code
 constexpr ResultCode ERR_NO_APPLICATION_AREA(ErrorModule::NFP, 152);
 } // namespace ErrCodes
 
 Module::Interface::Interface(std::shared_ptr<Module> module, Core::System& system, const char* name)
     : ServiceFramework(name), module(std::move(module)), system(system) {
     auto& kernel = system.Kernel();
-    nfc_tag_load = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                          "IUser:NFCTagDetected");
+    nfc_tag_load = Kernel::WritableEvent::CreateEventPair(kernel, "IUser:NFCTagDetected");
 }
 
 Module::Interface::~Interface() = default;
@@ -66,10 +62,9 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        deactivate_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IUser:DeactivateEvent");
-        availability_change_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic, "IUser:AvailabilityChangeEvent");
+        deactivate_event = Kernel::WritableEvent::CreateEventPair(kernel, "IUser:DeactivateEvent");
+        availability_change_event =
+            Kernel::WritableEvent::CreateEventPair(kernel, "IUser:AvailabilityChangeEvent");
     }
 
 private:

src/core/hle/service/nifm/nifm.cpp

@@ -69,10 +69,8 @@ public:
         RegisterHandlers(functions);
 
         auto& kernel = system.Kernel();
-        event1 = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                        "IRequest:Event1");
-        event2 = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
-                                                        "IRequest:Event2");
+        event1 = Kernel::WritableEvent::CreateEventPair(kernel, "IRequest:Event1");
+        event2 = Kernel::WritableEvent::CreateEventPair(kernel, "IRequest:Event2");
     }
 
 private:

src/core/hle/service/nim/nim.cpp

@@ -141,8 +141,7 @@ public:
 
         auto& kernel = system.Kernel();
         finished_event = Kernel::WritableEvent::CreateEventPair(
-            kernel, Kernel::ResetType::Automatic,
-            "IEnsureNetworkClockAvailabilityService:FinishEvent");
+            kernel, "IEnsureNetworkClockAvailabilityService:FinishEvent");
     }
 
 private:

src/core/hle/service/ns/ns.cpp

@@ -271,7 +271,7 @@ void IApplicationManagerInterface::GetApplicationControlData(Kernel::HLERequestC
                       "output buffer is too small! (actual={:016X}, expected_min=0x4000)", size);
             IPC::ResponseBuilder rb{ctx, 2};
             // TODO(DarkLordZach): Find a better error code for this.
-            rb.Push(ResultCode(-1));
+            rb.Push(RESULT_UNKNOWN);
             return;
         }
 
@@ -291,7 +291,7 @@ void IApplicationManagerInterface::GetApplicationControlData(Kernel::HLERequestC
                       0x4000 + control.second->GetSize());
             IPC::ResponseBuilder rb{ctx, 2};
             // TODO(DarkLordZach): Find a better error code for this.
-            rb.Push(ResultCode(-1));
+            rb.Push(RESULT_UNKNOWN);
             return;
         }
 

src/core/hle/service/ns/pl_u.cpp

@@ -97,7 +97,7 @@ void EncryptSharedFont(const std::vector<u32>& input, std::vector<u8>& output,
     const auto key = Common::swap32(EXPECTED_RESULT ^ EXPECTED_MAGIC);
     std::vector<u32> transformed_font(input.size() + 2);
     transformed_font[0] = Common::swap32(EXPECTED_MAGIC);
-    transformed_font[1] = Common::swap32(input.size() * sizeof(u32)) ^ key;
+    transformed_font[1] = Common::swap32(static_cast<u32>(input.size() * sizeof(u32))) ^ key;
     std::transform(input.begin(), input.end(), transformed_font.begin() + 2,
                    [key](u32 in) { return in ^ key; });
     std::memcpy(output.data() + offset, transformed_font.data(),

src/core/hle/service/nvdrv/nvdrv.cpp

@@ -40,8 +40,7 @@ Module::Module(Core::System& system) {
     auto& kernel = system.Kernel();
     for (u32 i = 0; i < MaxNvEvents; i++) {
         std::string event_label = fmt::format("NVDRV::NvEvent_{}", i);
-        events_interface.events[i] =
-            Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual, event_label);
+        events_interface.events[i] = Kernel::WritableEvent::CreateEventPair(kernel, event_label);
         events_interface.status[i] = EventState::Free;
         events_interface.registered[i] = false;
     }

src/core/hle/service/nvflinger/buffer_queue.cpp

@@ -16,8 +16,7 @@ namespace Service::NVFlinger {
 
 BufferQueue::BufferQueue(Kernel::KernelCore& kernel, u32 id, u64 layer_id)
     : id(id), layer_id(layer_id) {
-    buffer_wait_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                               "BufferQueue NativeHandle");
+    buffer_wait_event = Kernel::WritableEvent::CreateEventPair(kernel, "BufferQueue NativeHandle");
 }
 
 BufferQueue::~BufferQueue() = default;

src/core/hle/service/time/time.cpp

@@ -34,12 +34,12 @@ static void PosixToCalendar(u64 posix_time, CalendarTime& calendar_time,
         additional_info = {};
         return;
     }
-    calendar_time.year = tm->tm_year + 1900;
-    calendar_time.month = tm->tm_mon + 1;
-    calendar_time.day = tm->tm_mday;
-    calendar_time.hour = tm->tm_hour;
-    calendar_time.minute = tm->tm_min;
-    calendar_time.second = tm->tm_sec;
+    calendar_time.year = static_cast<u16_le>(tm->tm_year + 1900);
+    calendar_time.month = static_cast<u8>(tm->tm_mon + 1);
+    calendar_time.day = static_cast<u8>(tm->tm_mday);
+    calendar_time.hour = static_cast<u8>(tm->tm_hour);
+    calendar_time.minute = static_cast<u8>(tm->tm_min);
+    calendar_time.second = static_cast<u8>(tm->tm_sec);
 
     additional_info.day_of_week = tm->tm_wday;
     additional_info.day_of_year = tm->tm_yday;
@@ -322,7 +322,7 @@ void Module::Interface::GetClockSnapshot(Kernel::HLERequestContext& ctx) {
     if (tm == nullptr) {
         LOG_ERROR(Service_Time, "tm is a nullptr");
         IPC::ResponseBuilder rb{ctx, 2};
-        rb.Push(ResultCode(-1)); // TODO(ogniK): Find appropriate error code
+        rb.Push(RESULT_UNKNOWN); // TODO(ogniK): Find appropriate error code
         return;
     }
 
@@ -331,12 +331,12 @@ void Module::Interface::GetClockSnapshot(Kernel::HLERequestContext& ctx) {
     const SteadyClockTimePoint steady_clock_time_point{static_cast<u64_le>(ms.count() / 1000), {}};
 
     CalendarTime calendar_time{};
-    calendar_time.year = tm->tm_year + 1900;
-    calendar_time.month = tm->tm_mon + 1;
-    calendar_time.day = tm->tm_mday;
-    calendar_time.hour = tm->tm_hour;
-    calendar_time.minute = tm->tm_min;
-    calendar_time.second = tm->tm_sec;
+    calendar_time.year = static_cast<u16_le>(tm->tm_year + 1900);
+    calendar_time.month = static_cast<u8>(tm->tm_mon + 1);
+    calendar_time.day = static_cast<u8>(tm->tm_mday);
+    calendar_time.hour = static_cast<u8>(tm->tm_hour);
+    calendar_time.minute = static_cast<u8>(tm->tm_min);
+    calendar_time.second = static_cast<u8>(tm->tm_sec);
 
     ClockSnapshot clock_snapshot{};
     clock_snapshot.system_posix_time = time_since_epoch;

src/core/hle/service/vi/display/vi_display.cpp

@@ -17,8 +17,8 @@ namespace Service::VI {
 
 Display::Display(u64 id, std::string name, Core::System& system) : id{id}, name{std::move(name)} {
     auto& kernel = system.Kernel();
-    vsync_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Manual,
-                                                         fmt::format("Display VSync Event {}", id));
+    vsync_event =
+        Kernel::WritableEvent::CreateEventPair(kernel, fmt::format("Display VSync Event {}", id));
 }
 
 Display::~Display() = default;

src/core/hle/service/vi/vi.cpp

@@ -45,7 +45,7 @@ struct DisplayInfo {
 
     /// Whether or not the display has a limited number of layers.
     u8 has_limited_layers{1};
-    INSERT_PADDING_BYTES(7){};
+    INSERT_PADDING_BYTES(7);
 
     /// Indicates the total amount of layers supported by the display.
     /// @note This is only valid if has_limited_layers is set.
@@ -541,7 +541,7 @@ private:
             } else {
                 // Wait the current thread until a buffer becomes available
                 ctx.SleepClientThread(
-                    "IHOSBinderDriver::DequeueBuffer", -1,
+                    "IHOSBinderDriver::DequeueBuffer", UINT64_MAX,
                     [=](Kernel::SharedPtr<Kernel::Thread> thread, Kernel::HLERequestContext& ctx,
                         Kernel::ThreadWakeupReason reason) {
                         // Repeat TransactParcel DequeueBuffer when a buffer is available

src/core/perf_stats.cpp

@@ -81,7 +81,7 @@ double PerfStats::GetMeanFrametime() {
         return 0;
     }
     const double sum = std::accumulate(perf_history.begin() + IgnoreFrames,
-                                       perf_history.begin() + current_index, 0);
+                                       perf_history.begin() + current_index, 0.0);
     return sum / (current_index - IgnoreFrames);
 }
 

src/video_core/CMakeLists.txt

@@ -36,6 +36,8 @@ add_library(video_core STATIC
     memory_manager.h
     morton.cpp
     morton.h
+    rasterizer_accelerated.cpp
+    rasterizer_accelerated.h
     rasterizer_cache.cpp
     rasterizer_cache.h
     rasterizer_interface.h

src/video_core/buffer_cache/buffer_cache.h

@@ -12,6 +12,10 @@
 #include <utility>
 #include <vector>
 
+#include <boost/icl/interval_map.hpp>
+#include <boost/icl/interval_set.hpp>
+#include <boost/range/iterator_range.hpp>
+
 #include "common/alignment.h"
 #include "common/common_types.h"
 #include "core/core.h"
@@ -30,7 +34,7 @@ public:
     using BufferInfo = std::pair<const TBufferType*, u64>;
 
     BufferInfo UploadMemory(GPUVAddr gpu_addr, std::size_t size, std::size_t alignment = 4,
-                            bool is_written = false) {
+                            bool is_written = false, bool use_fast_cbuf = false) {
         std::lock_guard lock{mutex};
 
         auto& memory_manager = system.GPU().MemoryManager();
@@ -43,9 +47,13 @@ public:
         // Cache management is a big overhead, so only cache entries with a given size.
         // TODO: Figure out which size is the best for given games.
         constexpr std::size_t max_stream_size = 0x800;
-        if (size < max_stream_size) {
+        if (use_fast_cbuf || size < max_stream_size) {
             if (!is_written && !IsRegionWritten(cache_addr, cache_addr + size - 1)) {
-                return StreamBufferUpload(host_ptr, size, alignment);
+                if (use_fast_cbuf) {
+                    return ConstBufferUpload(host_ptr, size);
+                } else {
+                    return StreamBufferUpload(host_ptr, size, alignment);
+                }
             }
         }
 
@@ -152,6 +160,10 @@ protected:
     virtual void CopyBlock(const TBuffer& src, const TBuffer& dst, std::size_t src_offset,
                            std::size_t dst_offset, std::size_t size) = 0;
 
+    virtual BufferInfo ConstBufferUpload(const void* raw_pointer, std::size_t size) {
+        return {};
+    }
+
     /// Register an object into the cache
     void Register(const MapInterval& new_map, bool inherit_written = false) {
         const CacheAddr cache_ptr = new_map->GetStart();

src/video_core/engines/fermi_2d.cpp

@@ -28,6 +28,13 @@ void Fermi2D::CallMethod(const GPU::MethodCall& method_call) {
     }
 }
 
+std::pair<u32, u32> DelimitLine(u32 src_1, u32 src_2, u32 dst_1, u32 dst_2, u32 src_line) {
+    const u32 line_a = src_2 - src_1;
+    const u32 line_b = dst_2 - dst_1;
+    const u32 excess = std::max<s32>(0, line_a - src_line + src_1);
+    return {line_b - (excess * line_b) / line_a, excess};
+}
+
 void Fermi2D::HandleSurfaceCopy() {
     LOG_DEBUG(HW_GPU, "Requested a surface copy with operation {}",
               static_cast<u32>(regs.operation));
@@ -47,10 +54,27 @@ void Fermi2D::HandleSurfaceCopy() {
         src_blit_x2 = static_cast<u32>((regs.blit_src_x >> 32) + regs.blit_dst_width);
         src_blit_y2 = static_cast<u32>((regs.blit_src_y >> 32) + regs.blit_dst_height);
     }
+    u32 dst_blit_x2 = regs.blit_dst_x + regs.blit_dst_width;
+    u32 dst_blit_y2 = regs.blit_dst_y + regs.blit_dst_height;
+    const auto [new_dst_w, src_excess_x] =
+        DelimitLine(src_blit_x1, src_blit_x2, regs.blit_dst_x, dst_blit_x2, regs.src.width);
+    const auto [new_dst_h, src_excess_y] =
+        DelimitLine(src_blit_y1, src_blit_y2, regs.blit_dst_y, dst_blit_y2, regs.src.height);
+    dst_blit_x2 = new_dst_w + regs.blit_dst_x;
+    src_blit_x2 = src_blit_x2 - src_excess_x;
+    dst_blit_y2 = new_dst_h + regs.blit_dst_y;
+    src_blit_y2 = src_blit_y2 - src_excess_y;
+    const auto [new_src_w, dst_excess_x] =
+        DelimitLine(regs.blit_dst_x, dst_blit_x2, src_blit_x1, src_blit_x2, regs.dst.width);
+    const auto [new_src_h, dst_excess_y] =
+        DelimitLine(regs.blit_dst_y, dst_blit_y2, src_blit_y1, src_blit_y2, regs.dst.height);
+    src_blit_x2 = new_src_w + src_blit_x1;
+    dst_blit_x2 = dst_blit_x2 - dst_excess_x;
+    src_blit_y2 = new_src_h + src_blit_y1;
+    dst_blit_y2 = dst_blit_y2 - dst_excess_y;
     const Common::Rectangle<u32> src_rect{src_blit_x1, src_blit_y1, src_blit_x2, src_blit_y2};
-    const Common::Rectangle<u32> dst_rect{regs.blit_dst_x, regs.blit_dst_y,
-                                          regs.blit_dst_x + regs.blit_dst_width,
-                                          regs.blit_dst_y + regs.blit_dst_height};
+    const Common::Rectangle<u32> dst_rect{regs.blit_dst_x, regs.blit_dst_y, dst_blit_x2,
+                                          dst_blit_y2};
     Config copy_config;
     copy_config.operation = regs.operation;
     copy_config.filter = regs.blit_control.filter;

src/video_core/engines/fermi_2d.h

@@ -99,19 +99,19 @@ public:
 
         union {
             struct {
-                INSERT_PADDING_WORDS(0x80);
+                INSERT_UNION_PADDING_WORDS(0x80);
 
                 Surface dst;
 
-                INSERT_PADDING_WORDS(2);
+                INSERT_UNION_PADDING_WORDS(2);
 
                 Surface src;
 
-                INSERT_PADDING_WORDS(0x15);
+                INSERT_UNION_PADDING_WORDS(0x15);
 
                 Operation operation;
 
-                INSERT_PADDING_WORDS(0x177);
+                INSERT_UNION_PADDING_WORDS(0x177);
 
                 union {
                     u32 raw;
@@ -119,7 +119,7 @@ public:
                     BitField<4, 1, Filter> filter;
                 } blit_control;
 
-                INSERT_PADDING_WORDS(0x8);
+                INSERT_UNION_PADDING_WORDS(0x8);
 
                 u32 blit_dst_x;
                 u32 blit_dst_y;
@@ -130,7 +130,7 @@ public:
                 u64 blit_src_x;
                 u64 blit_src_y;
 
-                INSERT_PADDING_WORDS(0x21);
+                INSERT_UNION_PADDING_WORDS(0x21);
             };
             std::array<u32, NUM_REGS> reg_array;
         };

src/video_core/engines/kepler_compute.cpp

@@ -50,7 +50,7 @@ void KeplerCompute::CallMethod(const GPU::MethodCall& method_call) {
     }
 }
 
-Tegra::Texture::FullTextureInfo KeplerCompute::GetTexture(std::size_t offset) const {
+Texture::FullTextureInfo KeplerCompute::GetTexture(std::size_t offset) const {
     const std::bitset<8> cbuf_mask = launch_description.const_buffer_enable_mask.Value();
     ASSERT(cbuf_mask[regs.tex_cb_index]);
 
@@ -61,13 +61,11 @@ Tegra::Texture::FullTextureInfo KeplerCompute::GetTexture(std::size_t offset) co
     ASSERT(address < texinfo.Address() + texinfo.size);
 
     const Texture::TextureHandle tex_handle{memory_manager.Read<u32>(address)};
-    return GetTextureInfo(tex_handle, offset);
+    return GetTextureInfo(tex_handle);
 }
 
-Texture::FullTextureInfo KeplerCompute::GetTextureInfo(const Texture::TextureHandle tex_handle,
-                                                       std::size_t offset) const {
-    return Texture::FullTextureInfo{static_cast<u32>(offset), GetTICEntry(tex_handle.tic_id),
-                                    GetTSCEntry(tex_handle.tsc_id)};
+Texture::FullTextureInfo KeplerCompute::GetTextureInfo(Texture::TextureHandle tex_handle) const {
+    return Texture::FullTextureInfo{GetTICEntry(tex_handle.tic_id), GetTSCEntry(tex_handle.tsc_id)};
 }
 
 u32 KeplerCompute::AccessConstBuffer32(ShaderType stage, u64 const_buffer, u64 offset) const {
@@ -89,7 +87,7 @@ SamplerDescriptor KeplerCompute::AccessBindlessSampler(ShaderType stage, u64 con
     const GPUVAddr tex_info_address = tex_info_buffer.Address() + offset;
 
     const Texture::TextureHandle tex_handle{memory_manager.Read<u32>(tex_info_address)};
-    const Texture::FullTextureInfo tex_info = GetTextureInfo(tex_handle, offset);
+    const Texture::FullTextureInfo tex_info = GetTextureInfo(tex_handle);
     SamplerDescriptor result = SamplerDescriptor::FromTicTexture(tex_info.tic.texture_type.Value());
     result.is_shadow.Assign(tex_info.tsc.depth_compare_enabled.Value());
     return result;