loader/nsp: Move secondary loader initialization to constructor

Prevents nullptr bug when trying to dump the RomFS of an NSP resulting from secondary_loader not being initialized.

CMakeLists.txt

@@ -23,21 +23,21 @@ option(ENABLE_CUBEB "Enables the cubeb audio backend" ON)
 
 option(USE_DISCORD_PRESENCE "Enables Discord Rich Presence" OFF)
 
-if(NOT EXISTS ${CMAKE_SOURCE_DIR}/.git/hooks/pre-commit)
+if(NOT EXISTS ${PROJECT_SOURCE_DIR}/.git/hooks/pre-commit)
     message(STATUS "Copying pre-commit hook")
     file(COPY hooks/pre-commit
-        DESTINATION ${CMAKE_SOURCE_DIR}/.git/hooks)
+        DESTINATION ${PROJECT_SOURCE_DIR}/.git/hooks)
 endif()
 
 # Sanity check : Check that all submodules are present
 # =======================================================================
 
 function(check_submodules_present)
-    file(READ "${CMAKE_SOURCE_DIR}/.gitmodules" gitmodules)
+    file(READ "${PROJECT_SOURCE_DIR}/.gitmodules" gitmodules)
     string(REGEX MATCHALL "path *= *[^ \t\r\n]*" gitmodules ${gitmodules})
     foreach(module ${gitmodules})
         string(REGEX REPLACE "path *= *" "" module ${module})
-        if (NOT EXISTS "${CMAKE_SOURCE_DIR}/${module}/.git")
+        if (NOT EXISTS "${PROJECT_SOURCE_DIR}/${module}/.git")
             message(FATAL_ERROR "Git submodule ${module} not found. "
                     "Please run: git submodule update --init --recursive")
         endif()
@@ -45,17 +45,17 @@ function(check_submodules_present)
 endfunction()
 check_submodules_present()
 
-configure_file(${CMAKE_SOURCE_DIR}/dist/compatibility_list/compatibility_list.qrc
-               ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.qrc
+configure_file(${PROJECT_SOURCE_DIR}/dist/compatibility_list/compatibility_list.qrc
+               ${PROJECT_BINARY_DIR}/dist/compatibility_list/compatibility_list.qrc
                COPYONLY)
-if (ENABLE_COMPATIBILITY_LIST_DOWNLOAD AND NOT EXISTS ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
+if (ENABLE_COMPATIBILITY_LIST_DOWNLOAD AND NOT EXISTS ${PROJECT_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
     message(STATUS "Downloading compatibility list for yuzu...")
     file(DOWNLOAD
         https://api.yuzu-emu.org/gamedb/
-        "${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json" SHOW_PROGRESS)
+        "${PROJECT_BINARY_DIR}/dist/compatibility_list/compatibility_list.json" SHOW_PROGRESS)
 endif()
-if (NOT EXISTS ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
-    file(WRITE ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json "")
+if (NOT EXISTS ${PROJECT_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
+    file(WRITE ${PROJECT_BINARY_DIR}/dist/compatibility_list/compatibility_list.json "")
 endif()
 
 # Detect current compilation architecture and create standard definitions
@@ -170,7 +170,7 @@ endif()
 # On modern Unixes, this is typically already the case. The lone exception is
 # glibc, which may default to 32 bits. glibc allows this to be configured
 # by setting _FILE_OFFSET_BITS.
-if(CMAKE_SYSTEM_NAME STREQUAL "Linux")
+if(CMAKE_SYSTEM_NAME STREQUAL "Linux" OR MINGW)
     add_definitions(-D_FILE_OFFSET_BITS=64)
 endif()
 
@@ -178,10 +178,6 @@ endif()
 set_property(DIRECTORY APPEND PROPERTY
     COMPILE_DEFINITIONS $<$<CONFIG:Debug>:_DEBUG> $<$<NOT:$<CONFIG:Debug>>:NDEBUG>)
 
-
-math(EXPR EMU_ARCH_BITS ${CMAKE_SIZEOF_VOID_P}*8)
-add_definitions(-DEMU_ARCH_BITS=${EMU_ARCH_BITS})
-
 # System imported libraries
 # ======================
 
@@ -189,13 +185,13 @@ find_package(Boost 1.63.0 QUIET)
 if (NOT Boost_FOUND)
     message(STATUS "Boost 1.63.0 or newer not found, falling back to externals")
 
-    set(BOOST_ROOT "${CMAKE_SOURCE_DIR}/externals/boost")
+    set(BOOST_ROOT "${PROJECT_SOURCE_DIR}/externals/boost")
     set(Boost_NO_SYSTEM_PATHS OFF)
     find_package(Boost QUIET REQUIRED)
 endif()
 
 # Output binaries to bin/
-set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)
+set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${PROJECT_BINARY_DIR}/bin)
 
 # Prefer the -pthread flag on Linux.
 set(THREADS_PREFER_PTHREAD_FLAG ON)
@@ -264,7 +260,7 @@ if (YUZU_USE_BUNDLED_UNICORN)
         endif()
 
         set(UNICORN_FOUND YES)
-        set(UNICORN_PREFIX ${CMAKE_SOURCE_DIR}/externals/unicorn)
+        set(UNICORN_PREFIX ${PROJECT_SOURCE_DIR}/externals/unicorn)
         set(LIBUNICORN_LIBRARY "${UNICORN_PREFIX}/${UNICORN_LIB_NAME}" CACHE PATH "Path to Unicorn library" FORCE)
         set(LIBUNICORN_INCLUDE_DIR "${UNICORN_PREFIX}/include" CACHE PATH "Path to Unicorn headers" FORCE)
         set(UNICORN_DLL_DIR "${UNICORN_PREFIX}/" CACHE PATH "Path to unicorn dynamic library" FORCE)
@@ -356,12 +352,12 @@ set(CLANG_FORMAT_POSTFIX "-6.0")
 find_program(CLANG_FORMAT
     NAMES clang-format${CLANG_FORMAT_POSTFIX}
           clang-format
-    PATHS ${CMAKE_BINARY_DIR}/externals)
+    PATHS ${PROJECT_BINARY_DIR}/externals)
 # if find_program doesn't find it, try to download from externals
 if (NOT CLANG_FORMAT)
     if (WIN32)
         message(STATUS "Clang format not found! Downloading...")
-        set(CLANG_FORMAT "${CMAKE_BINARY_DIR}/externals/clang-format${CLANG_FORMAT_POSTFIX}.exe")
+        set(CLANG_FORMAT "${PROJECT_BINARY_DIR}/externals/clang-format${CLANG_FORMAT_POSTFIX}.exe")
         file(DOWNLOAD
             https://github.com/yuzu-emu/ext-windows-bin/raw/master/clang-format${CLANG_FORMAT_POSTFIX}.exe
             "${CLANG_FORMAT}" SHOW_PROGRESS
@@ -377,7 +373,7 @@ if (NOT CLANG_FORMAT)
 endif()
 
 if (CLANG_FORMAT)
-    set(SRCS ${CMAKE_SOURCE_DIR}/src)
+    set(SRCS ${PROJECT_SOURCE_DIR}/src)
     set(CCOMMENT "Running clang format against all the .h and .cpp files in src/")
     if (WIN32)
         add_custom_target(clang-format
@@ -450,10 +446,10 @@ endif()
 # http://standards.freedesktop.org/icon-theme-spec/icon-theme-spec-latest.html
 # http://standards.freedesktop.org/shared-mime-info-spec/shared-mime-info-spec-latest.html
 if(ENABLE_QT AND UNIX AND NOT APPLE)
-    install(FILES "${CMAKE_SOURCE_DIR}/dist/yuzu.desktop"
+    install(FILES "${PROJECT_SOURCE_DIR}/dist/yuzu.desktop"
             DESTINATION "${CMAKE_INSTALL_PREFIX}/share/applications")
-    install(FILES "${CMAKE_SOURCE_DIR}/dist/yuzu.svg"
+    install(FILES "${PROJECT_SOURCE_DIR}/dist/yuzu.svg"
             DESTINATION "${CMAKE_INSTALL_PREFIX}/share/icons/hicolor/scalable/apps")
-    install(FILES "${CMAKE_SOURCE_DIR}/dist/yuzu.xml"
+    install(FILES "${PROJECT_SOURCE_DIR}/dist/yuzu.xml"
             DESTINATION "${CMAKE_INSTALL_PREFIX}/share/mime/packages")
 endif()

src/audio_core/time_stretch.cpp

@@ -10,8 +10,7 @@
 
 namespace AudioCore {
 
-TimeStretcher::TimeStretcher(u32 sample_rate, u32 channel_count)
-    : m_sample_rate(sample_rate), m_channel_count(channel_count) {
+TimeStretcher::TimeStretcher(u32 sample_rate, u32 channel_count) : m_sample_rate{sample_rate} {
     m_sound_touch.setChannels(channel_count);
     m_sound_touch.setSampleRate(sample_rate);
     m_sound_touch.setPitch(1.0);

src/audio_core/time_stretch.h

@@ -27,7 +27,6 @@ public:
 
 private:
     u32 m_sample_rate;
-    u32 m_channel_count;
     soundtouch::SoundTouch m_sound_touch;
     double m_stretch_ratio = 1.0;
 };

src/common/CMakeLists.txt

@@ -64,8 +64,6 @@ add_library(common STATIC
     logging/text_formatter.cpp
     logging/text_formatter.h
     math_util.h
-    memory_util.cpp
-    memory_util.h
     microprofile.cpp
     microprofile.h
     microprofileui.h

src/common/file_util.cpp

@@ -15,21 +15,24 @@
 #ifdef _WIN32
 #include <windows.h>
 // windows.h needs to be included before other windows headers
-#include <commdlg.h> // for GetSaveFileName
-#include <direct.h>  // getcwd
+#include <direct.h> // getcwd
 #include <io.h>
 #include <shellapi.h>
 #include <shlobj.h> // for SHGetFolderPath
 #include <tchar.h>
 #include "common/string_util.h"
 
-// 64 bit offsets for windows
+#ifdef _MSC_VER
+// 64 bit offsets for MSVC
 #define fseeko _fseeki64
 #define ftello _ftelli64
-#define atoll _atoi64
+#define fileno _fileno
+#endif
+
+// 64 bit offsets for MSVC and MinGW. MinGW also needs this for using _wstat64
 #define stat _stat64
 #define fstat _fstat64
-#define fileno _fileno
+
 #else
 #ifdef __APPLE__
 #include <sys/param.h>

src/common/logging/backend.cpp

@@ -196,6 +196,7 @@ void FileBackend::Write(const Entry& entry) {
     SUB(Service, NFP)                                                                              \
     SUB(Service, NIFM)                                                                             \
     SUB(Service, NIM)                                                                              \
+    SUB(Service, NPNS)                                                                             \
     SUB(Service, NS)                                                                               \
     SUB(Service, NVDRV)                                                                            \
     SUB(Service, PCIE)                                                                             \
@@ -204,10 +205,12 @@ void FileBackend::Write(const Entry& entry) {
     SUB(Service, PM)                                                                               \
     SUB(Service, PREPO)                                                                            \
     SUB(Service, PSC)                                                                              \
+    SUB(Service, PSM)                                                                              \
     SUB(Service, SET)                                                                              \
     SUB(Service, SM)                                                                               \
     SUB(Service, SPL)                                                                              \
     SUB(Service, SSL)                                                                              \
+    SUB(Service, TCAP)                                                                             \
     SUB(Service, Time)                                                                             \
     SUB(Service, USB)                                                                              \
     SUB(Service, VI)                                                                               \

src/common/logging/log.h

@@ -83,6 +83,7 @@ enum class Class : ClassType {
     Service_NFP,       ///< The NFP service
     Service_NIFM,      ///< The NIFM (Network interface) service
     Service_NIM,       ///< The NIM service
+    Service_NPNS,      ///< The NPNS service
     Service_NS,        ///< The NS services
     Service_NVDRV,     ///< The NVDRV (Nvidia driver) service
     Service_PCIE,      ///< The PCIe service
@@ -96,6 +97,7 @@ enum class Class : ClassType {
     Service_SM,        ///< The SM (Service manager) service
     Service_SPL,       ///< The SPL service
     Service_SSL,       ///< The SSL service
+    Service_TCAP,      ///< The TCAP service.
     Service_Time,      ///< The time service
     Service_USB,       ///< The USB (Universal Serial Bus) service
     Service_VI,        ///< The VI (Video interface) service

src/common/memory_util.cpp

@@ -1,177 +0,0 @@
-// Copyright 2013 Dolphin Emulator Project / 2014 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#include "common/logging/log.h"
-#include "common/memory_util.h"
-
-#ifdef _WIN32
-#include <windows.h>
-// Windows.h needs to be included before psapi.h
-#include <psapi.h>
-#include "common/common_funcs.h"
-#include "common/string_util.h"
-#else
-#include <cstdlib>
-#include <sys/mman.h>
-#endif
-
-#if !defined(_WIN32) && defined(ARCHITECTURE_x86_64) && !defined(MAP_32BIT)
-#include <unistd.h>
-#define PAGE_MASK (getpagesize() - 1)
-#define round_page(x) ((((unsigned long)(x)) + PAGE_MASK) & ~(PAGE_MASK))
-#endif
-
-// This is purposely not a full wrapper for virtualalloc/mmap, but it
-// provides exactly the primitive operations that Dolphin needs.
-
-void* AllocateExecutableMemory(std::size_t size, bool low) {
-#if defined(_WIN32)
-    void* ptr = VirtualAlloc(nullptr, size, MEM_COMMIT, PAGE_EXECUTE_READWRITE);
-#else
-    static char* map_hint = nullptr;
-#if defined(ARCHITECTURE_x86_64) && !defined(MAP_32BIT)
-    // This OS has no flag to enforce allocation below the 4 GB boundary,
-    // but if we hint that we want a low address it is very likely we will
-    // get one.
-    // An older version of this code used MAP_FIXED, but that has the side
-    // effect of discarding already mapped pages that happen to be in the
-    // requested virtual memory range (such as the emulated RAM, sometimes).
-    if (low && (!map_hint))
-        map_hint = (char*)round_page(512 * 1024 * 1024); /* 0.5 GB rounded up to the next page */
-#endif
-    void* ptr = mmap(map_hint, size, PROT_READ | PROT_WRITE | PROT_EXEC,
-                     MAP_ANON | MAP_PRIVATE
-#if defined(ARCHITECTURE_x86_64) && defined(MAP_32BIT)
-                         | (low ? MAP_32BIT : 0)
-#endif
-                         ,
-                     -1, 0);
-#endif /* defined(_WIN32) */
-
-#ifdef _WIN32
-    if (ptr == nullptr) {
-#else
-    if (ptr == MAP_FAILED) {
-        ptr = nullptr;
-#endif
-        LOG_ERROR(Common_Memory, "Failed to allocate executable memory");
-    }
-#if !defined(_WIN32) && defined(ARCHITECTURE_x86_64) && !defined(MAP_32BIT)
-    else {
-        if (low) {
-            map_hint += size;
-            map_hint = (char*)round_page(map_hint); /* round up to the next page */
-        }
-    }
-#endif
-
-#if EMU_ARCH_BITS == 64
-    if ((u64)ptr >= 0x80000000 && low == true)
-        LOG_ERROR(Common_Memory, "Executable memory ended up above 2GB!");
-#endif
-
-    return ptr;
-}
-
-void* AllocateMemoryPages(std::size_t size) {
-#ifdef _WIN32
-    void* ptr = VirtualAlloc(nullptr, size, MEM_COMMIT, PAGE_READWRITE);
-#else
-    void* ptr = mmap(nullptr, size, PROT_READ | PROT_WRITE, MAP_ANON | MAP_PRIVATE, -1, 0);
-
-    if (ptr == MAP_FAILED)
-        ptr = nullptr;
-#endif
-
-    if (ptr == nullptr)
-        LOG_ERROR(Common_Memory, "Failed to allocate raw memory");
-
-    return ptr;
-}
-
-void* AllocateAlignedMemory(std::size_t size, std::size_t alignment) {
-#ifdef _WIN32
-    void* ptr = _aligned_malloc(size, alignment);
-#else
-    void* ptr = nullptr;
-#ifdef ANDROID
-    ptr = memalign(alignment, size);
-#else
-    if (posix_memalign(&ptr, alignment, size) != 0)
-        LOG_ERROR(Common_Memory, "Failed to allocate aligned memory");
-#endif
-#endif
-
-    if (ptr == nullptr)
-        LOG_ERROR(Common_Memory, "Failed to allocate aligned memory");
-
-    return ptr;
-}
-
-void FreeMemoryPages(void* ptr, std::size_t size) {
-    if (ptr) {
-#ifdef _WIN32
-        if (!VirtualFree(ptr, 0, MEM_RELEASE))
-            LOG_ERROR(Common_Memory, "FreeMemoryPages failed!\n{}", GetLastErrorMsg());
-#else
-        munmap(ptr, size);
-#endif
-    }
-}
-
-void FreeAlignedMemory(void* ptr) {
-    if (ptr) {
-#ifdef _WIN32
-        _aligned_free(ptr);
-#else
-        free(ptr);
-#endif
-    }
-}
-
-void WriteProtectMemory(void* ptr, std::size_t size, bool allowExecute) {
-#ifdef _WIN32
-    DWORD oldValue;
-    if (!VirtualProtect(ptr, size, allowExecute ? PAGE_EXECUTE_READ : PAGE_READONLY, &oldValue))
-        LOG_ERROR(Common_Memory, "WriteProtectMemory failed!\n{}", GetLastErrorMsg());
-#else
-    mprotect(ptr, size, allowExecute ? (PROT_READ | PROT_EXEC) : PROT_READ);
-#endif
-}
-
-void UnWriteProtectMemory(void* ptr, std::size_t size, bool allowExecute) {
-#ifdef _WIN32
-    DWORD oldValue;
-    if (!VirtualProtect(ptr, size, allowExecute ? PAGE_EXECUTE_READWRITE : PAGE_READWRITE,
-                        &oldValue))
-        LOG_ERROR(Common_Memory, "UnWriteProtectMemory failed!\n{}", GetLastErrorMsg());
-#else
-    mprotect(ptr, size,
-             allowExecute ? (PROT_READ | PROT_WRITE | PROT_EXEC) : PROT_WRITE | PROT_READ);
-#endif
-}
-
-std::string MemUsage() {
-#ifdef _WIN32
-#pragma comment(lib, "psapi")
-    DWORD processID = GetCurrentProcessId();
-    HANDLE hProcess;
-    PROCESS_MEMORY_COUNTERS pmc;
-    std::string Ret;
-
-    // Print information about the memory usage of the process.
-
-    hProcess = OpenProcess(PROCESS_QUERY_INFORMATION | PROCESS_VM_READ, FALSE, processID);
-    if (nullptr == hProcess)
-        return "MemUsage Error";
-
-    if (GetProcessMemoryInfo(hProcess, &pmc, sizeof(pmc)))
-        Ret = fmt::format("{} K", Common::ThousandSeparate(pmc.WorkingSetSize / 1024, 7));
-
-    CloseHandle(hProcess);
-    return Ret;
-#else
-    return "";
-#endif
-}

src/common/memory_util.h

@@ -1,21 +0,0 @@
-// Copyright 2013 Dolphin Emulator Project / 2014 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <cstddef>
-#include <string>
-
-void* AllocateExecutableMemory(std::size_t size, bool low = true);
-void* AllocateMemoryPages(std::size_t size);
-void FreeMemoryPages(void* ptr, std::size_t size);
-void* AllocateAlignedMemory(std::size_t size, std::size_t alignment);
-void FreeAlignedMemory(void* ptr);
-void WriteProtectMemory(void* ptr, std::size_t size, bool executable = false);
-void UnWriteProtectMemory(void* ptr, std::size_t size, bool allowExecute = false);
-std::string MemUsage();
-
-inline int GetPageSize() {
-    return 4096;
-}

src/core/CMakeLists.txt

@@ -156,6 +156,8 @@ add_library(core STATIC
     hle/service/am/omm.h
     hle/service/am/spsm.cpp
     hle/service/am/spsm.h
+    hle/service/am/tcap.cpp
+    hle/service/am/tcap.h
     hle/service/aoc/aoc_u.cpp
     hle/service/aoc/aoc_u.h
     hle/service/apm/apm.cpp
@@ -280,6 +282,8 @@ add_library(core STATIC
     hle/service/nifm/nifm.h
     hle/service/nim/nim.cpp
     hle/service/nim/nim.h
+    hle/service/npns/npns.cpp
+    hle/service/npns/npns.h
     hle/service/ns/ns.cpp
     hle/service/ns/ns.h
     hle/service/ns/pl_u.cpp

src/core/crypto/key_manager.cpp

@@ -713,7 +713,6 @@ void KeyManager::DeriveBase() {
 
     const auto sbk = GetKey(S128KeyType::SecureBoot);
     const auto tsec = GetKey(S128KeyType::TSEC);
-    const auto master_source = GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::Master));
 
     for (size_t i = 0; i < revisions.size(); ++i) {
         if (!revisions[i])

src/core/file_sys/card_image.cpp

@@ -168,10 +168,6 @@ VirtualDir XCI::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
-bool XCI::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
-
 Loader::ResultStatus XCI::AddNCAFromPartition(XCIPartition part) {
     if (partitions[static_cast<std::size_t>(part)] == nullptr) {
         return Loader::ResultStatus::ErrorXCIMissingPartition;

src/core/file_sys/card_image.h

@@ -94,9 +94,6 @@ public:
 
     VirtualDir GetParentDirectory() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     Loader::ResultStatus AddNCAFromPartition(XCIPartition part);
 

src/core/file_sys/content_archive.cpp

@@ -546,7 +546,4 @@ u64 NCA::GetBaseIVFCOffset() const {
     return ivfc_offset;
 }
 
-bool NCA::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys

src/core/file_sys/content_archive.h

@@ -100,9 +100,6 @@ public:
     // Returns the base ivfc offset used in BKTR patching.
     u64 GetBaseIVFCOffset() const;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     bool CheckSupportedNCA(const NCAHeader& header);
     bool HandlePotentialHeaderDecryption();

src/core/file_sys/ips_layer.cpp

@@ -99,7 +99,7 @@ VirtualFile PatchIPS(const VirtualFile& in, const VirtualFile& ips) {
             u16 rle_size{};
             if (ips->ReadObject(&rle_size, offset) != sizeof(u16))
                 return nullptr;
-            rle_size = Common::swap16(data_size);
+            rle_size = Common::swap16(rle_size);
             offset += sizeof(u16);
 
             const auto data = ips->ReadByte(offset++);

src/core/file_sys/partition_filesystem.cpp

@@ -83,7 +83,7 @@ std::vector<std::shared_ptr<VfsFile>> PartitionFilesystem::GetFiles() const {
 }
 
 std::vector<std::shared_ptr<VfsDirectory>> PartitionFilesystem::GetSubdirectories() const {
-    return pfs_dirs;
+    return {};
 }
 
 std::string PartitionFilesystem::GetName() const {
@@ -103,18 +103,4 @@ void PartitionFilesystem::PrintDebugInfo() const {
                   pfs_files[i]->GetName(), pfs_files[i]->GetSize());
     }
 }
-
-bool PartitionFilesystem::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    const auto iter = std::find(pfs_files.begin(), pfs_files.end(), file);
-    if (iter == pfs_files.end())
-        return false;
-
-    const std::ptrdiff_t offset = std::distance(pfs_files.begin(), iter);
-    pfs_files[offset] = std::move(pfs_files.back());
-    pfs_files.pop_back();
-
-    pfs_dirs.emplace_back(std::move(dir));
-
-    return true;
-}
 } // namespace FileSys

src/core/file_sys/partition_filesystem.h

@@ -35,9 +35,6 @@ public:
     std::shared_ptr<VfsDirectory> GetParentDirectory() const override;
     void PrintDebugInfo() const;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     struct Header {
         u32_le magic;
@@ -84,7 +81,6 @@ private:
     std::size_t content_offset = 0;
 
     std::vector<VirtualFile> pfs_files;
-    std::vector<VirtualDir> pfs_dirs;
 };
 
 } // namespace FileSys

src/core/file_sys/patch_manager.cpp

@@ -168,7 +168,8 @@ bool PatchManager::HasNSOPatch(const std::array<u8, 32>& build_id_) const {
 
 static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType type) {
     const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
-    if (type != ContentRecordType::Program || load_dir == nullptr || load_dir->GetSize() <= 0) {
+    if ((type != ContentRecordType::Program && type != ContentRecordType::Data) ||
+        load_dir == nullptr || load_dir->GetSize() <= 0) {
         return;
     }
 
@@ -218,7 +219,7 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content
                                         title_id, static_cast<u8>(type))
                                 .c_str();
 
-    if (type == ContentRecordType::Program)
+    if (type == ContentRecordType::Program || type == ContentRecordType::Data)
         LOG_INFO(Loader, log_string);
     else
         LOG_DEBUG(Loader, log_string);

src/core/file_sys/registered_cache.cpp

@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
 #include <regex>
 #include <mbedtls/sha256.h>
 #include "common/assert.h"
@@ -30,6 +31,14 @@ bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs)
     return (lhs.title_id < rhs.title_id) || (lhs.title_id == rhs.title_id && lhs.type < rhs.type);
 }
 
+bool operator==(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs) {
+    return std::tie(lhs.title_id, lhs.type) == std::tie(rhs.title_id, rhs.type);
+}
+
+bool operator!=(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs) {
+    return !operator==(lhs, rhs);
+}
+
 static bool FollowsTwoDigitDirFormat(std::string_view name) {
     static const std::regex two_digit_regex("000000[0-9A-F]{2}", std::regex_constants::ECMAScript |
                                                                      std::regex_constants::icase);
@@ -593,6 +602,9 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntries() const {
             },
             [](const CNMT& c, const ContentRecord& r) { return true; });
     }
+
+    std::sort(out.begin(), out.end());
+    out.erase(std::unique(out.begin(), out.end()), out.end());
     return out;
 }
 
@@ -616,6 +628,9 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntriesFilter(
                 return true;
             });
     }
+
+    std::sort(out.begin(), out.end());
+    out.erase(std::unique(out.begin(), out.end()), out.end());
     return out;
 }
 } // namespace FileSys

src/core/file_sys/registered_cache.h

@@ -50,6 +50,10 @@ constexpr u64 GetUpdateTitleID(u64 base_title_id) {
 // boost flat_map requires operator< for O(log(n)) lookups.
 bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs);
 
+// std unique requires operator== to identify duplicates.
+bool operator==(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs);
+bool operator!=(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs);
+
 /*
  * A class that catalogues NCAs in the registered directory structure.
  * Nintendo's registered format follows this structure:
@@ -60,8 +64,8 @@ bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs)
  *         | 00
  *         | 01 <- Actual content split along 4GB boundaries. (optional)
  *
- * (This impl also supports substituting the nca dir for an nca file, as that's more convenient when
- * 4GB splitting can be ignored.)
+ * (This impl also supports substituting the nca dir for an nca file, as that's more convenient
+ * when 4GB splitting can be ignored.)
  */
 class RegisteredCache {
     friend class RegisteredCacheUnion;

src/core/file_sys/submission_package.cpp

@@ -205,10 +205,6 @@ VirtualDir NSP::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
-bool NSP::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
-
 void NSP::InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files) {
     exefs = pfs;
 

src/core/file_sys/submission_package.h

@@ -55,9 +55,6 @@ public:
 
     VirtualDir GetParentDirectory() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     void InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files);
     void ReadNCAs(const std::vector<VirtualFile>& files);

src/core/file_sys/vfs.cpp

@@ -472,10 +472,14 @@ bool VfsRawCopy(const VirtualFile& src, const VirtualFile& dest, std::size_t blo
     std::vector<u8> temp(std::min(block_size, src->GetSize()));
     for (std::size_t i = 0; i < src->GetSize(); i += block_size) {
         const auto read = std::min(block_size, src->GetSize() - i);
-        const auto block = src->Read(temp.data(), read, i);
 
-        if (dest->Write(temp.data(), read, i) != read)
+        if (src->Read(temp.data(), read, i) != read) {
             return false;
+        }
+
+        if (dest->Write(temp.data(), read, i) != read) {
+            return false;
+        }
     }
 
     return true;

src/core/file_sys/vfs.h

@@ -262,36 +262,8 @@ public:
     // item name -> type.
     virtual std::map<std::string, VfsEntryType, std::less<>> GetEntries() const;
 
-    // Interprets the file with name file instead as a directory of type directory.
-    // The directory must have a constructor that takes a single argument of type
-    // std::shared_ptr<VfsFile>. Allows to reinterpret container files (i.e NCA, zip, XCI, etc) as a
-    // subdirectory in one call.
-    template <typename Directory>
-    bool InterpretAsDirectory(std::string_view file) {
-        auto file_p = GetFile(file);
-
-        if (file_p == nullptr) {
-            return false;
-        }
-
-        return ReplaceFileWithSubdirectory(file_p, std::make_shared<Directory>(file_p));
-    }
-
-    bool InterpretAsDirectory(const std::function<VirtualDir(VirtualFile)>& function,
-                              const std::string& file) {
-        auto file_p = GetFile(file);
-        if (file_p == nullptr)
-            return false;
-        return ReplaceFileWithSubdirectory(file_p, function(file_p));
-    }
-
     // Returns the full path of this directory as a string, recursively
     virtual std::string GetFullPath() const;
-
-protected:
-    // Backend for InterpretAsDirectory.
-    // Removes all references to file and adds a reference to dir in the directory's implementation.
-    virtual bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) = 0;
 };
 
 // A convenience partial-implementation of VfsDirectory that stubs out methods that should only work

src/core/file_sys/vfs_layered.cpp

@@ -126,7 +126,4 @@ bool LayeredVfsDirectory::Rename(std::string_view name_) {
     return true;
 }
 
-bool LayeredVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys

src/core/file_sys/vfs_layered.h

@@ -39,9 +39,6 @@ public:
     bool DeleteFile(std::string_view name) override;
     bool Rename(std::string_view name) override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     std::vector<VirtualDir> dirs;
     std::string name;

src/core/file_sys/vfs_real.cpp

@@ -430,7 +430,4 @@ std::map<std::string, VfsEntryType, std::less<>> RealVfsDirectory::GetEntries()
     return out;
 }
 
-bool RealVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys

src/core/file_sys/vfs_real.h

@@ -100,9 +100,6 @@ public:
     std::string GetFullPath() const override;
     std::map<std::string, VfsEntryType, std::less<>> GetEntries() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     RealVfsDirectory(RealVfsFilesystem& base, const std::string& path, Mode perms = Mode::Read);
 

src/core/file_sys/vfs_vector.cpp

@@ -132,11 +132,4 @@ void VectorVfsDirectory::AddFile(VirtualFile file) {
 void VectorVfsDirectory::AddDirectory(VirtualDir dir) {
     dirs.push_back(std::move(dir));
 }
-
-bool VectorVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    if (!DeleteFile(file->GetName()))
-        return false;
-    dirs.emplace_back(std::move(dir));
-    return true;
-}
 } // namespace FileSys

src/core/file_sys/vfs_vector.h

@@ -57,9 +57,6 @@ public:
     virtual void AddFile(VirtualFile file);
     virtual void AddDirectory(VirtualDir dir);
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     std::vector<VirtualFile> files;
     std::vector<VirtualDir> dirs;

src/core/file_sys/xts_archive.cpp

@@ -163,7 +163,4 @@ std::shared_ptr<VfsDirectory> NAX::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
-bool NAX::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
-    return false;
-}
 } // namespace FileSys

src/core/file_sys/xts_archive.h

@@ -51,9 +51,6 @@ public:
 
     std::shared_ptr<VfsDirectory> GetParentDirectory() const override;
 
-protected:
-    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
-
 private:
     Loader::ResultStatus Parse(std::string_view path);
 

src/core/hle/ipc.h

@@ -14,11 +14,6 @@ namespace IPC {
 /// Size of the command buffer area, in 32-bit words.
 constexpr std::size_t COMMAND_BUFFER_LENGTH = 0x100 / sizeof(u32);
 
-// These errors are commonly returned by invalid IPC translations, so alias them here for
-// convenience.
-// TODO(yuriks): These will probably go away once translation is implemented inside the kernel.
-constexpr auto ERR_INVALID_HANDLE = Kernel::ERR_INVALID_HANDLE_OS;
-
 enum class ControlCommand : u32 {
     ConvertSessionToDomain = 0,
     ConvertDomainToSession = 1,

src/core/hle/kernel/errors.h

@@ -10,11 +10,6 @@ namespace Kernel {
 
 namespace ErrCodes {
 enum {
-    // TODO(Subv): Remove these 3DS OS error codes.
-    SessionClosedByRemote = 26,
-    NoPendingSessions = 35,
-    InvalidBufferDescriptor = 48,
-
     // Confirmed Switch OS error codes
     MaxConnectionsReached = 7,
     InvalidSize = 101,
@@ -26,6 +21,7 @@ enum {
     InvalidThreadPriority = 112,
     InvalidProcessorId = 113,
     InvalidHandle = 114,
+    InvalidPointer = 115,
     InvalidCombination = 116,
     Timeout = 117,
     SynchronizationCanceled = 118,
@@ -33,6 +29,7 @@ enum {
     InvalidEnumValue = 120,
     NoSuchEntry = 121,
     AlreadyRegistered = 122,
+    SessionClosed = 123,
     InvalidState = 125,
     ResourceLimitExceeded = 132,
 };
@@ -41,18 +38,14 @@ enum {
 // WARNING: The kernel is quite inconsistent in it's usage of errors code. Make sure to always
 // double check that the code matches before re-using the constant.
 
-// TODO(bunnei): Replace -1 with correct errors for Switch OS
 constexpr ResultCode ERR_HANDLE_TABLE_FULL(ErrorModule::Kernel, ErrCodes::HandleTableFull);
-constexpr ResultCode ERR_SESSION_CLOSED_BY_REMOTE(-1);
+constexpr ResultCode ERR_SESSION_CLOSED_BY_REMOTE(ErrorModule::Kernel, ErrCodes::SessionClosed);
 constexpr ResultCode ERR_PORT_NAME_TOO_LONG(ErrorModule::Kernel, ErrCodes::TooLarge);
 constexpr ResultCode ERR_MAX_CONNECTIONS_REACHED(ErrorModule::Kernel,
                                                  ErrCodes::MaxConnectionsReached);
 constexpr ResultCode ERR_INVALID_ENUM_VALUE(ErrorModule::Kernel, ErrCodes::InvalidEnumValue);
-constexpr ResultCode ERR_INVALID_ENUM_VALUE_FND(-1);
-constexpr ResultCode ERR_INVALID_COMBINATION(-1);
 constexpr ResultCode ERR_INVALID_COMBINATION_KERNEL(ErrorModule::Kernel,
                                                     ErrCodes::InvalidCombination);
-constexpr ResultCode ERR_OUT_OF_MEMORY(-1);
 constexpr ResultCode ERR_INVALID_ADDRESS(ErrorModule::Kernel, ErrCodes::InvalidAddress);
 constexpr ResultCode ERR_INVALID_ADDRESS_STATE(ErrorModule::Kernel, ErrCodes::InvalidMemoryState);
 constexpr ResultCode ERR_INVALID_MEMORY_PERMISSIONS(ErrorModule::Kernel,
@@ -65,14 +58,8 @@ constexpr ResultCode ERR_ALREADY_REGISTERED(ErrorModule::Kernel, ErrCodes::Alrea
 constexpr ResultCode ERR_INVALID_STATE(ErrorModule::Kernel, ErrCodes::InvalidState);
 constexpr ResultCode ERR_INVALID_THREAD_PRIORITY(ErrorModule::Kernel,
                                                  ErrCodes::InvalidThreadPriority);
-constexpr ResultCode ERR_INVALID_POINTER(-1);
-constexpr ResultCode ERR_INVALID_OBJECT_ADDR(-1);
-constexpr ResultCode ERR_NOT_AUTHORIZED(-1);
-/// Alternate code returned instead of ERR_INVALID_HANDLE in some code paths.
-constexpr ResultCode ERR_INVALID_HANDLE_OS(-1);
+constexpr ResultCode ERR_INVALID_POINTER(ErrorModule::Kernel, ErrCodes::InvalidPointer);
 constexpr ResultCode ERR_NOT_FOUND(ErrorModule::Kernel, ErrCodes::NoSuchEntry);
 constexpr ResultCode RESULT_TIMEOUT(ErrorModule::Kernel, ErrCodes::Timeout);
-/// Returned when Accept() is called on a port with no sessions to be accepted.
-constexpr ResultCode ERR_NO_PENDING_SESSIONS(-1);
 
 } // namespace Kernel

src/core/hle/kernel/hle_ipc.cpp

@@ -77,7 +77,8 @@ HLERequestContext::HLERequestContext(SharedPtr<Kernel::ServerSession> server_ses
 
 HLERequestContext::~HLERequestContext() = default;
 
-void HLERequestContext::ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming) {
+void HLERequestContext::ParseCommandBuffer(const HandleTable& handle_table, u32_le* src_cmdbuf,
+                                           bool incoming) {
     IPC::RequestParser rp(src_cmdbuf);
     command_header = std::make_shared<IPC::CommandHeader>(rp.PopRaw<IPC::CommandHeader>());
 
@@ -94,8 +95,6 @@ void HLERequestContext::ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming) {
             rp.Skip(2, false);
         }
         if (incoming) {
-            auto& handle_table = Core::System::GetInstance().Kernel().HandleTable();
-
             // Populate the object lists with the data in the IPC request.
             for (u32 handle = 0; handle < handle_descriptor_header->num_handles_to_copy; ++handle) {
                 copy_objects.push_back(handle_table.GetGeneric(rp.Pop<Handle>()));
@@ -189,10 +188,9 @@ void HLERequestContext::ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming) {
     rp.Skip(1, false); // The command is actually an u64, but we don't use the high part.
 }
 
-ResultCode HLERequestContext::PopulateFromIncomingCommandBuffer(u32_le* src_cmdbuf,
-                                                                Process& src_process,
-                                                                HandleTable& src_table) {
-    ParseCommandBuffer(src_cmdbuf, true);
+ResultCode HLERequestContext::PopulateFromIncomingCommandBuffer(const HandleTable& handle_table,
+                                                                u32_le* src_cmdbuf) {
+    ParseCommandBuffer(handle_table, src_cmdbuf, true);
     if (command_header->type == IPC::CommandType::Close) {
         // Close does not populate the rest of the IPC header
         return RESULT_SUCCESS;
@@ -207,14 +205,17 @@ ResultCode HLERequestContext::PopulateFromIncomingCommandBuffer(u32_le* src_cmdb
     return RESULT_SUCCESS;
 }
 
-ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(const Thread& thread) {
+ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(Thread& thread) {
+    auto& owner_process = *thread.GetOwnerProcess();
+    auto& handle_table = owner_process.GetHandleTable();
+
     std::array<u32, IPC::COMMAND_BUFFER_LENGTH> dst_cmdbuf;
-    Memory::ReadBlock(*thread.GetOwnerProcess(), thread.GetTLSAddress(), dst_cmdbuf.data(),
+    Memory::ReadBlock(owner_process, thread.GetTLSAddress(), dst_cmdbuf.data(),
                       dst_cmdbuf.size() * sizeof(u32));
 
     // The header was already built in the internal command buffer. Attempt to parse it to verify
     // the integrity and then copy it over to the target command buffer.
-    ParseCommandBuffer(cmd_buf.data(), false);
+    ParseCommandBuffer(handle_table, cmd_buf.data(), false);
 
     // The data_size already includes the payload header, the padding and the domain header.
     std::size_t size = data_payload_offset + command_header->data_size -
@@ -236,8 +237,6 @@ ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(const Thread& thread)
         ASSERT(copy_objects.size() == handle_descriptor_header->num_handles_to_copy);
         ASSERT(move_objects.size() == handle_descriptor_header->num_handles_to_move);
 
-        auto& handle_table = Core::System::GetInstance().Kernel().HandleTable();
-
         // We don't make a distinction between copy and move handles when translating since HLE
         // services don't deal with handles directly. However, the guest applications might check
         // for specific values in each of these descriptors.
@@ -268,7 +267,7 @@ ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(const Thread& thread)
     }
 
     // Copy the translated command buffer back into the thread's command buffer area.
-    Memory::WriteBlock(*thread.GetOwnerProcess(), thread.GetTLSAddress(), dst_cmdbuf.data(),
+    Memory::WriteBlock(owner_process, thread.GetTLSAddress(), dst_cmdbuf.data(),
                        dst_cmdbuf.size() * sizeof(u32));
 
     return RESULT_SUCCESS;

src/core/hle/kernel/hle_ipc.h

@@ -24,10 +24,10 @@ class ServiceFrameworkBase;
 namespace Kernel {
 
 class Domain;
+class Event;
 class HandleTable;
 class HLERequestContext;
 class Process;
-class Event;
 
 /**
  * Interface implemented by HLE Session handlers.
@@ -126,13 +126,12 @@ public:
                                        u64 timeout, WakeupCallback&& callback,
                                        Kernel::SharedPtr<Kernel::Event> event = nullptr);
 
-    void ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming);
-
     /// Populates this context with data from the requesting process/thread.
-    ResultCode PopulateFromIncomingCommandBuffer(u32_le* src_cmdbuf, Process& src_process,
-                                                 HandleTable& src_table);
+    ResultCode PopulateFromIncomingCommandBuffer(const HandleTable& handle_table,
+                                                 u32_le* src_cmdbuf);
+
     /// Writes data from this context back to the requesting process/thread.
-    ResultCode WriteToOutgoingCommandBuffer(const Thread& thread);
+    ResultCode WriteToOutgoingCommandBuffer(Thread& thread);
 
     u32_le GetCommand() const {
         return command;
@@ -255,6 +254,8 @@ public:
     std::string Description() const;
 
 private:
+    void ParseCommandBuffer(const HandleTable& handle_table, u32_le* src_cmdbuf, bool incoming);
+
     std::array<u32, IPC::COMMAND_BUFFER_LENGTH> cmd_buf;
     SharedPtr<Kernel::ServerSession> server_session;
     // TODO(yuriks): Check common usage of this and optimize size accordingly

src/core/hle/kernel/kernel.cpp

@@ -118,7 +118,6 @@ struct KernelCore::Impl {
         process_list.clear();
         current_process = nullptr;
 
-        handle_table.Clear();
         resource_limits.fill(nullptr);
 
         thread_wakeup_callback_handle_table.Clear();
@@ -209,7 +208,6 @@ struct KernelCore::Impl {
     std::vector<SharedPtr<Process>> process_list;
     Process* current_process = nullptr;
 
-    Kernel::HandleTable handle_table;
     std::array<SharedPtr<ResourceLimit>, 4> resource_limits;
 
     /// The event type of the generic timer callback event
@@ -241,14 +239,6 @@ void KernelCore::Shutdown() {
     impl->Shutdown();
 }
 
-Kernel::HandleTable& KernelCore::HandleTable() {
-    return impl->handle_table;
-}
-
-const Kernel::HandleTable& KernelCore::HandleTable() const {
-    return impl->handle_table;
-}
-
 SharedPtr<ResourceLimit> KernelCore::ResourceLimitForCategory(
     ResourceLimitCategory category) const {
     return impl->resource_limits.at(static_cast<std::size_t>(category));

src/core/hle/kernel/kernel.h

@@ -47,12 +47,6 @@ public:
     /// Clears all resources in use by the kernel instance.
     void Shutdown();
 
-    /// Provides a reference to the handle table.
-    Kernel::HandleTable& HandleTable();
-
-    /// Provides a const reference to the handle table.
-    const Kernel::HandleTable& HandleTable() const;
-
     /// Retrieves a shared pointer to a ResourceLimit identified by the given category.
     SharedPtr<ResourceLimit> ResourceLimitForCategory(ResourceLimitCategory category) const;
 

src/core/hle/kernel/process.cpp

@@ -232,6 +232,12 @@ void Process::LoadModule(CodeSet module_, VAddr base_addr) {
     MapSegment(module_.CodeSegment(), VMAPermission::ReadExecute, MemoryState::CodeStatic);
     MapSegment(module_.RODataSegment(), VMAPermission::Read, MemoryState::CodeMutable);
     MapSegment(module_.DataSegment(), VMAPermission::ReadWrite, MemoryState::CodeMutable);
+
+    // Clear instruction cache in CPU JIT
+    Core::System::GetInstance().ArmInterface(0).ClearInstructionCache();
+    Core::System::GetInstance().ArmInterface(1).ClearInstructionCache();
+    Core::System::GetInstance().ArmInterface(2).ClearInstructionCache();
+    Core::System::GetInstance().ArmInterface(3).ClearInstructionCache();
 }
 
 ResultVal<VAddr> Process::HeapAllocate(VAddr target, u64 size, VMAPermission perms) {

src/core/hle/kernel/process.h

@@ -13,6 +13,7 @@
 #include <boost/container/static_vector.hpp>
 #include "common/bit_field.h"
 #include "common/common_types.h"
+#include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/vm_manager.h"
@@ -142,6 +143,16 @@ public:
         return vm_manager;
     }
 
+    /// Gets a reference to the process' handle table.
+    HandleTable& GetHandleTable() {
+        return handle_table;
+    }
+
+    /// Gets a const reference to the process' handle table.
+    const HandleTable& GetHandleTable() const {
+        return handle_table;
+    }
+
     /// Gets the current status of the process
     ProcessStatus GetStatus() const {
         return status;
@@ -294,6 +305,9 @@ private:
     /// specified by metadata provided to the process during loading.
     bool is_64bit_process = true;
 
+    /// Per-process handle table for storing created object handles in.
+    HandleTable handle_table;
+
     std::string name;
 };
 

src/core/hle/kernel/server_port.cpp

@@ -18,7 +18,7 @@ ServerPort::~ServerPort() = default;
 
 ResultVal<SharedPtr<ServerSession>> ServerPort::Accept() {
     if (pending_sessions.empty()) {
-        return ERR_NO_PENDING_SESSIONS;
+        return ERR_NOT_FOUND;
     }
 
     auto session = std::move(pending_sessions.back());
@@ -28,7 +28,7 @@ ResultVal<SharedPtr<ServerSession>> ServerPort::Accept() {
 
 bool ServerPort::ShouldWait(Thread* thread) const {
     // If there are no pending sessions, we wait until a new one is added.
-    return pending_sessions.size() == 0;
+    return pending_sessions.empty();
 }
 
 void ServerPort::Acquire(Thread* thread) {

src/core/hle/kernel/server_session.cpp

@@ -107,8 +107,7 @@ ResultCode ServerSession::HandleSyncRequest(SharedPtr<Thread> thread) {
     // similar.
     Kernel::HLERequestContext context(this);
     u32* cmd_buf = (u32*)Memory::GetPointer(thread->GetTLSAddress());
-    context.PopulateFromIncomingCommandBuffer(cmd_buf, *Core::CurrentProcess(),
-                                              kernel.HandleTable());
+    context.PopulateFromIncomingCommandBuffer(kernel.CurrentProcess()->GetHandleTable(), cmd_buf);
 
     ResultCode result = RESULT_SUCCESS;
     // If the session has been converted to a domain, handle the domain request

src/core/hle/kernel/shared_memory.cpp

@@ -80,20 +80,19 @@ SharedPtr<SharedMemory> SharedMemory::CreateForApplet(
 
 ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermission permissions,
                              MemoryPermission other_permissions) {
-
-    MemoryPermission own_other_permissions =
+    const MemoryPermission own_other_permissions =
         target_process == owner_process ? this->permissions : this->other_permissions;
 
     // Automatically allocated memory blocks can only be mapped with other_permissions = DontCare
     if (base_address == 0 && other_permissions != MemoryPermission::DontCare) {
-        return ERR_INVALID_COMBINATION;
+        return ERR_INVALID_MEMORY_PERMISSIONS;
     }
 
     // Error out if the requested permissions don't match what the creator process allows.
     if (static_cast<u32>(permissions) & ~static_cast<u32>(own_other_permissions)) {
         LOG_ERROR(Kernel, "cannot map id={}, address=0x{:X} name={}, permissions don't match",
                   GetObjectId(), address, name);
-        return ERR_INVALID_COMBINATION;
+        return ERR_INVALID_MEMORY_PERMISSIONS;
     }
 
     // Error out if the provided permissions are not compatible with what the creator process needs.

src/core/hle/kernel/svc.cpp

@@ -189,14 +189,15 @@ static ResultCode ConnectToNamedPort(Handle* out_handle, VAddr port_name_address
     CASCADE_RESULT(client_session, client_port->Connect());
 
     // Return the client session
-    CASCADE_RESULT(*out_handle, kernel.HandleTable().Create(client_session));
+    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    CASCADE_RESULT(*out_handle, handle_table.Create(client_session));
     return RESULT_SUCCESS;
 }
 
 /// Makes a blocking IPC call to an OS service.
 static ResultCode SendSyncRequest(Handle handle) {
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<ClientSession> session = kernel.HandleTable().Get<ClientSession>(handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    SharedPtr<ClientSession> session = handle_table.Get<ClientSession>(handle);
     if (!session) {
         LOG_ERROR(Kernel_SVC, "called with invalid handle=0x{:08X}", handle);
         return ERR_INVALID_HANDLE;
@@ -215,8 +216,8 @@ static ResultCode SendSyncRequest(Handle handle) {
 static ResultCode GetThreadId(u32* thread_id, Handle thread_handle) {
     LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -229,8 +230,8 @@ static ResultCode GetThreadId(u32* thread_id, Handle thread_handle) {
 static ResultCode GetProcessId(u32* process_id, Handle process_handle) {
     LOG_TRACE(Kernel_SVC, "called process=0x{:08X}", process_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Process> process = kernel.HandleTable().Get<Process>(process_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Process> process = handle_table.Get<Process>(process_handle);
     if (!process) {
         return ERR_INVALID_HANDLE;
     }
@@ -273,11 +274,11 @@ static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64
 
     using ObjectPtr = Thread::ThreadWaitObjects::value_type;
     Thread::ThreadWaitObjects objects(handle_count);
-    auto& kernel = Core::System::GetInstance().Kernel();
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
 
     for (u64 i = 0; i < handle_count; ++i) {
         const Handle handle = Memory::Read32(handles_address + i * sizeof(Handle));
-        const auto object = kernel.HandleTable().Get<WaitObject>(handle);
+        const auto object = handle_table.Get<WaitObject>(handle);
 
         if (object == nullptr) {
             return ERR_INVALID_HANDLE;
@@ -325,8 +326,8 @@ static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64
 static ResultCode CancelSynchronization(Handle thread_handle) {
     LOG_TRACE(Kernel_SVC, "called thread=0x{:X}", thread_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -354,7 +355,7 @@ static ResultCode ArbitrateLock(Handle holding_thread_handle, VAddr mutex_addr,
         return ERR_INVALID_ADDRESS;
     }
 
-    auto& handle_table = Core::System::GetInstance().Kernel().HandleTable();
+    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
     return Mutex::TryAcquire(handle_table, mutex_addr, holding_thread_handle,
                              requesting_thread_handle);
 }
@@ -374,9 +375,19 @@ static ResultCode ArbitrateUnlock(VAddr mutex_addr) {
     return Mutex::Release(mutex_addr);
 }
 
+enum class BreakType : u32 {
+    Panic = 0,
+    AssertionFailed = 1,
+    PreNROLoad = 3,
+    PostNROLoad = 4,
+    PreNROUnload = 5,
+    PostNROUnload = 6,
+};
+
 struct BreakReason {
     union {
         u32 raw;
+        BitField<0, 30, BreakType> break_type;
         BitField<31, 1, u32> signal_debugger;
     };
 };
@@ -384,12 +395,48 @@ struct BreakReason {
 /// Break program execution
 static void Break(u32 reason, u64 info1, u64 info2) {
     BreakReason break_reason{reason};
-    if (break_reason.signal_debugger) {
-        LOG_ERROR(
+
+    switch (break_reason.break_type) {
+    case BreakType::Panic:
+        LOG_CRITICAL(Debug_Emulated, "Signalling debugger, PANIC! info1=0x{:016X}, info2=0x{:016X}",
+                     info1, info2);
+        break;
+    case BreakType::AssertionFailed:
+        LOG_CRITICAL(Debug_Emulated,
+                     "Signalling debugger, Assertion failed! info1=0x{:016X}, info2=0x{:016X}",
+                     info1, info2);
+        break;
+    case BreakType::PreNROLoad:
+        LOG_WARNING(
             Debug_Emulated,
-            "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}",
-            reason, info1, info2);
-    } else {
+            "Signalling debugger, Attempting to load an NRO at 0x{:016X} with size 0x{:016X}",
+            info1, info2);
+        break;
+    case BreakType::PostNROLoad:
+        LOG_WARNING(Debug_Emulated,
+                    "Signalling debugger, Loaded an NRO at 0x{:016X} with size 0x{:016X}", info1,
+                    info2);
+        break;
+    case BreakType::PreNROUnload:
+        LOG_WARNING(
+            Debug_Emulated,
+            "Signalling debugger, Attempting to unload an NRO at 0x{:016X} with size 0x{:016X}",
+            info1, info2);
+        break;
+    case BreakType::PostNROUnload:
+        LOG_WARNING(Debug_Emulated,
+                    "Signalling debugger, Unloaded an NRO at 0x{:016X} with size 0x{:016X}", info1,
+                    info2);
+        break;
+    default:
+        LOG_WARNING(
+            Debug_Emulated,
+            "Signalling debugger, Unknown break reason {}, info1=0x{:016X}, info2=0x{:016X}",
+            static_cast<u32>(break_reason.break_type.Value()), info1, info2);
+        break;
+    }
+
+    if (!break_reason.signal_debugger) {
         LOG_CRITICAL(
             Debug_Emulated,
             "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}",
@@ -499,13 +546,12 @@ static ResultCode SetThreadActivity(Handle handle, u32 unknown) {
 static ResultCode GetThreadContext(VAddr thread_context, Handle handle) {
     LOG_DEBUG(Kernel_SVC, "called, context=0x{:08X}, thread=0x{:X}", thread_context, handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(handle);
+    const auto* current_process = Core::CurrentProcess();
+    const SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
 
-    const auto* current_process = Core::CurrentProcess();
     if (thread->GetOwnerProcess() != current_process) {
         return ERR_INVALID_HANDLE;
     }
@@ -531,10 +577,11 @@ static ResultCode GetThreadContext(VAddr thread_context, Handle handle) {
 
 /// Gets the priority for the specified thread
 static ResultCode GetThreadPriority(u32* priority, Handle handle) {
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(handle);
-    if (!thread)
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(handle);
+    if (!thread) {
         return ERR_INVALID_HANDLE;
+    }
 
     *priority = thread->GetPriority();
     return RESULT_SUCCESS;
@@ -546,16 +593,18 @@ static ResultCode SetThreadPriority(Handle handle, u32 priority) {
         return ERR_INVALID_THREAD_PRIORITY;
     }
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(handle);
-    if (!thread)
-        return ERR_INVALID_HANDLE;
+    const auto* const current_process = Core::CurrentProcess();
 
     // Note: The kernel uses the current process's resource limit instead of
     // the one from the thread owner's resource limit.
-    const ResourceLimit& resource_limit = Core::CurrentProcess()->GetResourceLimit();
+    const ResourceLimit& resource_limit = current_process->GetResourceLimit();
     if (resource_limit.GetMaxResourceValue(ResourceType::Priority) > priority) {
-        return ERR_NOT_AUTHORIZED;
+        return ERR_INVALID_THREAD_PRIORITY;
+    }
+
+    SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle);
+    if (!thread) {
+        return ERR_INVALID_HANDLE;
     }
 
     thread->SetPriority(priority);
@@ -595,15 +644,13 @@ static ResultCode MapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 s
         return ERR_INVALID_MEMORY_PERMISSIONS;
     }
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    auto shared_memory = kernel.HandleTable().Get<SharedMemory>(shared_memory_handle);
+    auto* const current_process = Core::CurrentProcess();
+    auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle);
     if (!shared_memory) {
         return ERR_INVALID_HANDLE;
     }
 
-    auto* const current_process = Core::CurrentProcess();
     const auto& vm_manager = current_process->VMManager();
-
     if (!vm_manager.IsWithinASLRRegion(addr, size)) {
         return ERR_INVALID_MEMORY_RANGE;
     }
@@ -627,15 +674,13 @@ static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64
         return ERR_INVALID_ADDRESS_STATE;
     }
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    auto shared_memory = kernel.HandleTable().Get<SharedMemory>(shared_memory_handle);
+    auto* const current_process = Core::CurrentProcess();
+    auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle);
     if (!shared_memory) {
         return ERR_INVALID_HANDLE;
     }
 
-    auto* const current_process = Core::CurrentProcess();
     const auto& vm_manager = current_process->VMManager();
-
     if (!vm_manager.IsWithinASLRRegion(addr, size)) {
         return ERR_INVALID_MEMORY_RANGE;
     }
@@ -646,15 +691,14 @@ static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64
 /// Query process memory
 static ResultCode QueryProcessMemory(MemoryInfo* memory_info, PageInfo* /*page_info*/,
                                      Handle process_handle, u64 addr) {
-
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<Process> process = kernel.HandleTable().Get<Process>(process_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    SharedPtr<Process> process = handle_table.Get<Process>(process_handle);
     if (!process) {
         return ERR_INVALID_HANDLE;
     }
     auto vma = process->VMManager().FindVMA(addr);
     memory_info->attributes = 0;
-    if (vma == Core::CurrentProcess()->VMManager().vma_map.end()) {
+    if (vma == process->VMManager().vma_map.end()) {
         memory_info->base_address = 0;
         memory_info->permission = static_cast<u32>(VMAPermission::None);
         memory_info->size = 0;
@@ -695,20 +739,19 @@ static void ExitProcess() {
 /// Creates a new thread
 static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, VAddr stack_top,
                                u32 priority, s32 processor_id) {
-    std::string name = fmt::format("thread-{:X}", entry_point);
-
     if (priority > THREADPRIO_LOWEST) {
         return ERR_INVALID_THREAD_PRIORITY;
     }
 
-    const ResourceLimit& resource_limit = Core::CurrentProcess()->GetResourceLimit();
+    auto* const current_process = Core::CurrentProcess();
+    const ResourceLimit& resource_limit = current_process->GetResourceLimit();
     if (resource_limit.GetMaxResourceValue(ResourceType::Priority) > priority) {
-        return ERR_NOT_AUTHORIZED;
+        return ERR_INVALID_THREAD_PRIORITY;
     }
 
     if (processor_id == THREADPROCESSORID_DEFAULT) {
         // Set the target CPU to the one specified in the process' exheader.
-        processor_id = Core::CurrentProcess()->GetDefaultProcessorID();
+        processor_id = current_process->GetDefaultProcessorID();
         ASSERT(processor_id != THREADPROCESSORID_DEFAULT);
     }
 
@@ -723,11 +766,13 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
         return ERR_INVALID_PROCESSOR_ID;
     }
 
+    const std::string name = fmt::format("thread-{:X}", entry_point);
     auto& kernel = Core::System::GetInstance().Kernel();
     CASCADE_RESULT(SharedPtr<Thread> thread,
                    Thread::Create(kernel, name, entry_point, priority, arg, processor_id, stack_top,
-                                  *Core::CurrentProcess()));
-    const auto new_guest_handle = kernel.HandleTable().Create(thread);
+                                  *current_process));
+
+    const auto new_guest_handle = current_process->GetHandleTable().Create(thread);
     if (new_guest_handle.Failed()) {
         return new_guest_handle.Code();
     }
@@ -748,8 +793,8 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
 static ResultCode StartThread(Handle thread_handle) {
     LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -796,8 +841,8 @@ static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_var
         "called mutex_addr={:X}, condition_variable_addr={:X}, thread_handle=0x{:08X}, timeout={}",
         mutex_addr, condition_variable_addr, thread_handle, nano_seconds);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     ASSERT(thread);
 
     CASCADE_CODE(Mutex::Release(mutex_addr));
@@ -908,9 +953,9 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target
                                                mutex_val | Mutex::MutexHasWaitersFlag));
 
             // The mutex is already owned by some other thread, make this thread wait on it.
-            auto& kernel = Core::System::GetInstance().Kernel();
-            Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask);
-            auto owner = kernel.HandleTable().Get<Thread>(owner_handle);
+            const Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask);
+            const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+            auto owner = handle_table.Get<Thread>(owner_handle);
             ASSERT(owner);
             ASSERT(thread->GetStatus() == ThreadStatus::WaitMutex);
             thread->InvalidateWakeupCallback();
@@ -989,16 +1034,16 @@ static u64 GetSystemTick() {
 static ResultCode CloseHandle(Handle handle) {
     LOG_TRACE(Kernel_SVC, "Closing handle 0x{:08X}", handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    return kernel.HandleTable().Close(handle);
+    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    return handle_table.Close(handle);
 }
 
 /// Reset an event
 static ResultCode ResetSignal(Handle handle) {
     LOG_WARNING(Kernel_SVC, "(STUBBED) called handle 0x{:08X}", handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    auto event = kernel.HandleTable().Get<Event>(handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    auto event = handle_table.Get<Event>(handle);
 
     ASSERT(event != nullptr);
 
@@ -1017,8 +1062,8 @@ static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32
 static ResultCode GetThreadCoreMask(Handle thread_handle, u32* core, u64* mask) {
     LOG_TRACE(Kernel_SVC, "called, handle=0x{:08X}", thread_handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -1033,8 +1078,8 @@ static ResultCode SetThreadCoreMask(Handle thread_handle, u32 core, u64 mask) {
     LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, mask=0x{:16X}, core=0x{:X}", thread_handle,
               mask, core);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    const SharedPtr<Thread> thread = kernel.HandleTable().Get<Thread>(thread_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
     if (!thread) {
         return ERR_INVALID_HANDLE;
     }
@@ -1095,7 +1140,7 @@ static ResultCode CreateSharedMemory(Handle* handle, u64 size, u32 local_permiss
     }
 
     auto& kernel = Core::System::GetInstance().Kernel();
-    auto& handle_table = kernel.HandleTable();
+    auto& handle_table = Core::CurrentProcess()->GetHandleTable();
     auto shared_mem_handle =
         SharedMemory::Create(kernel, handle_table.Get<Process>(KernelHandle::CurrentProcess), size,
                              local_perms, remote_perms);
@@ -1107,10 +1152,12 @@ static ResultCode CreateSharedMemory(Handle* handle, u64 size, u32 local_permiss
 static ResultCode ClearEvent(Handle handle) {
     LOG_TRACE(Kernel_SVC, "called, event=0x{:08X}", handle);
 
-    auto& kernel = Core::System::GetInstance().Kernel();
-    SharedPtr<Event> evt = kernel.HandleTable().Get<Event>(handle);
-    if (evt == nullptr)
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    SharedPtr<Event> evt = handle_table.Get<Event>(handle);
+    if (evt == nullptr) {
         return ERR_INVALID_HANDLE;
+    }
+
     evt->Clear();
     return RESULT_SUCCESS;
 }
@@ -1123,8 +1170,8 @@ static ResultCode GetProcessInfo(u64* out, Handle process_handle, u32 type) {
         Status,
     };
 
-    const auto& kernel = Core::System::GetInstance().Kernel();
-    const auto process = kernel.HandleTable().Get<Process>(process_handle);
+    const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+    const auto process = handle_table.Get<Process>(process_handle);
     if (!process) {
         return ERR_INVALID_HANDLE;
     }

src/core/hle/kernel/thread.cpp

@@ -266,7 +266,7 @@ SharedPtr<Thread> SetupMainThread(KernelCore& kernel, VAddr entry_point, u32 pri
     SharedPtr<Thread> thread = std::move(thread_res).Unwrap();
 
     // Register 1 must be a handle to the main thread
-    const Handle guest_handle = kernel.HandleTable().Create(thread).Unwrap();
+    const Handle guest_handle = owner_process.GetHandleTable().Create(thread).Unwrap();
     thread->SetGuestHandle(guest_handle);
     thread->GetContext().cpu_registers[1] = guest_handle;
 

src/core/hle/kernel/vm_manager.cpp

@@ -143,6 +143,26 @@ ResultVal<VMManager::VMAHandle> VMManager::MapBackingMemory(VAddr target, u8* me
     return MakeResult<VMAHandle>(MergeAdjacent(vma_handle));
 }
 
+ResultVal<VAddr> VMManager::FindFreeRegion(u64 size) const {
+    // Find the first Free VMA.
+    const VAddr base = GetASLRRegionBaseAddress();
+    const VMAHandle vma_handle = std::find_if(vma_map.begin(), vma_map.end(), [&](const auto& vma) {
+        if (vma.second.type != VMAType::Free)
+            return false;
+
+        const VAddr vma_end = vma.second.base + vma.second.size;
+        return vma_end > base && vma_end >= base + size;
+    });
+
+    if (vma_handle == vma_map.end()) {
+        // TODO(Subv): Find the correct error code here.
+        return ResultCode(-1);
+    }
+
+    const VAddr target = std::max(base, vma_handle->second.base);
+    return MakeResult<VAddr>(target);
+}
+
 ResultVal<VMManager::VMAHandle> VMManager::MapMMIO(VAddr target, PAddr paddr, u64 size,
                                                    MemoryState state,
                                                    Memory::MemoryHookPointer mmio_handler) {

src/core/hle/kernel/vm_manager.h

@@ -157,6 +157,14 @@ public:
      */
     ResultVal<VMAHandle> MapBackingMemory(VAddr target, u8* memory, u64 size, MemoryState state);
 
+    /**
+     * Finds the first free address that can hold a region of the desired size.
+     *
+     * @param size Size of the desired region.
+     * @return The found free address.
+     */
+    ResultVal<VAddr> FindFreeRegion(u64 size) const;
+
     /**
      * Maps a memory-mapped IO region at a given address.
      *

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

@@ -2,9 +2,13 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
 #include <array>
+#include "common/common_paths.h"
 #include "common/common_types.h"
+#include "common/file_util.h"
 #include "common/logging/log.h"
+#include "common/string_util.h"
 #include "common/swap.h"
 #include "core/core_timing.h"
 #include "core/hle/ipc_helpers.h"
@@ -16,6 +20,7 @@
 #include "core/hle/service/acc/profile_manager.h"
 
 namespace Service::Account {
+
 // TODO: RE this structure
 struct UserData {
     INSERT_PADDING_WORDS(1);
@@ -27,6 +32,29 @@ struct UserData {
 };
 static_assert(sizeof(UserData) == 0x80, "UserData structure has incorrect size");
 
+// Smallest JPEG https://github.com/mathiasbynens/small/blob/master/jpeg.jpg
+// used as a backup should the one on disk not exist
+constexpr u32 backup_jpeg_size = 107;
+constexpr std::array<u8, backup_jpeg_size> backup_jpeg{{
+    0xff, 0xd8, 0xff, 0xdb, 0x00, 0x43, 0x00, 0x03, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x02, 0x02,
+    0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x06, 0x04, 0x04, 0x04, 0x04, 0x04, 0x08, 0x06, 0x06, 0x05,
+    0x06, 0x09, 0x08, 0x0a, 0x0a, 0x09, 0x08, 0x09, 0x09, 0x0a, 0x0c, 0x0f, 0x0c, 0x0a, 0x0b, 0x0e,
+    0x0b, 0x09, 0x09, 0x0d, 0x11, 0x0d, 0x0e, 0x0f, 0x10, 0x10, 0x11, 0x10, 0x0a, 0x0c, 0x12, 0x13,
+    0x12, 0x10, 0x13, 0x0f, 0x10, 0x10, 0x10, 0xff, 0xc9, 0x00, 0x0b, 0x08, 0x00, 0x01, 0x00, 0x01,
+    0x01, 0x01, 0x11, 0x00, 0xff, 0xcc, 0x00, 0x06, 0x00, 0x10, 0x10, 0x05, 0xff, 0xda, 0x00, 0x08,
+    0x01, 0x01, 0x00, 0x00, 0x3f, 0x00, 0xd2, 0xcf, 0x20, 0xff, 0xd9,
+}};
+
+static std::string GetImagePath(UUID uuid) {
+    return FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+           "/system/save/8000000000000010/su/avators/" + uuid.FormatSwitch() + ".jpg";
+}
+
+static constexpr u32 SanitizeJPEGSize(std::size_t size) {
+    constexpr std::size_t max_jpeg_image_size = 0x20000;
+    return static_cast<u32>(std::min(size, max_jpeg_image_size));
+}
+
 class IProfile final : public ServiceFramework<IProfile> {
 public:
     explicit IProfile(UUID user_id, ProfileManager& profile_manager)
@@ -73,32 +101,42 @@ private:
     }
 
     void LoadImage(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_ACC, "(STUBBED) called");
-        // smallest jpeg https://github.com/mathiasbynens/small/blob/master/jpeg.jpg
-        // TODO(mailwl): load actual profile image from disk, width 256px, max size 0x20000
-        constexpr u32 jpeg_size = 107;
-        static constexpr std::array<u8, jpeg_size> jpeg{
-            0xff, 0xd8, 0xff, 0xdb, 0x00, 0x43, 0x00, 0x03, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03,
-            0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x06, 0x04, 0x04, 0x04, 0x04, 0x04,
-            0x08, 0x06, 0x06, 0x05, 0x06, 0x09, 0x08, 0x0a, 0x0a, 0x09, 0x08, 0x09, 0x09, 0x0a,
-            0x0c, 0x0f, 0x0c, 0x0a, 0x0b, 0x0e, 0x0b, 0x09, 0x09, 0x0d, 0x11, 0x0d, 0x0e, 0x0f,
-            0x10, 0x10, 0x11, 0x10, 0x0a, 0x0c, 0x12, 0x13, 0x12, 0x10, 0x13, 0x0f, 0x10, 0x10,
-            0x10, 0xff, 0xc9, 0x00, 0x0b, 0x08, 0x00, 0x01, 0x00, 0x01, 0x01, 0x01, 0x11, 0x00,
-            0xff, 0xcc, 0x00, 0x06, 0x00, 0x10, 0x10, 0x05, 0xff, 0xda, 0x00, 0x08, 0x01, 0x01,
-            0x00, 0x00, 0x3f, 0x00, 0xd2, 0xcf, 0x20, 0xff, 0xd9,
-        };
-        ctx.WriteBuffer(jpeg);
+        LOG_DEBUG(Service_ACC, "called");
+
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(jpeg_size);
+
+        const FileUtil::IOFile image(GetImagePath(user_id), "rb");
+        if (!image.IsOpen()) {
+            LOG_WARNING(Service_ACC,
+                        "Failed to load user provided image! Falling back to built-in backup...");
+            ctx.WriteBuffer(backup_jpeg);
+            rb.Push<u32>(backup_jpeg_size);
+            return;
+        }
+
+        const u32 size = SanitizeJPEGSize(image.GetSize());
+        std::vector<u8> buffer(size);
+        image.ReadBytes(buffer.data(), buffer.size());
+
+        ctx.WriteBuffer(buffer.data(), buffer.size());
+        rb.Push<u32>(size);
     }
 
     void GetImageSize(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_ACC, "(STUBBED) called");
-        constexpr u32 jpeg_size = 107;
+        LOG_DEBUG(Service_ACC, "called");
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(jpeg_size);
+
+        const FileUtil::IOFile image(GetImagePath(user_id), "rb");
+
+        if (!image.IsOpen()) {
+            LOG_WARNING(Service_ACC,
+                        "Failed to load user provided image! Falling back to built-in backup...");
+            rb.Push<u32>(backup_jpeg_size);
+        } else {
+            rb.Push<u32>(SanitizeJPEGSize(image.GetSize()));
+        }
     }
 
     const ProfileManager& profile_manager;

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

@@ -3,41 +3,66 @@
 // Refer to the license.txt file included.
 
 #include <random>
-#include <boost/optional.hpp>
+
+#include "common/file_util.h"
 #include "core/hle/service/acc/profile_manager.h"
 #include "core/settings.h"
 
 namespace Service::Account {
+
+struct UserRaw {
+    UUID uuid;
+    UUID uuid2;
+    u64 timestamp;
+    ProfileUsername username;
+    INSERT_PADDING_BYTES(0x80);
+};
+static_assert(sizeof(UserRaw) == 0xC8, "UserRaw has incorrect size.");
+
+struct ProfileDataRaw {
+    INSERT_PADDING_BYTES(0x10);
+    std::array<UserRaw, MAX_USERS> users;
+};
+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_ARGUMENT_IS_NULL(ErrorModule::Account, 20);
 
-const UUID& UUID::Generate() {
+constexpr char ACC_SAVE_AVATORS_BASE_PATH[] = "/system/save/8000000000000010/su/avators/";
+
+UUID UUID::Generate() {
     std::random_device device;
     std::mt19937 gen(device());
     std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max());
-    uuid[0] = distribution(gen);
-    uuid[1] = distribution(gen);
-    return *this;
+    return UUID{distribution(gen), distribution(gen)};
 }
 
 ProfileManager::ProfileManager() {
-    // TODO(ogniK): Create the default user we have for now until loading/saving users is added
-    auto user_uuid = UUID{1, 0};
-    ASSERT(CreateNewUser(user_uuid, Settings::values.username).IsSuccess());
-    OpenUser(user_uuid);
+    ParseUserSaveFile();
+
+    if (user_count == 0)
+        CreateNewUser(UUID::Generate(), "yuzu");
+
+    auto current = std::clamp<int>(Settings::values.current_user, 0, MAX_USERS - 1);
+    if (UserExistsIndex(current))
+        current = 0;
+
+    OpenUser(*GetUser(current));
 }
 
-ProfileManager::~ProfileManager() = default;
+ProfileManager::~ProfileManager() {
+    WriteUserSaveFile();
+}
 
 /// After a users creation it needs to be "registered" to the system. AddToProfiles handles the
 /// internal management of the users profiles
-boost::optional<std::size_t> ProfileManager::AddToProfiles(const ProfileInfo& user) {
+std::optional<std::size_t> ProfileManager::AddToProfiles(const ProfileInfo& profile) {
     if (user_count >= MAX_USERS) {
-        return boost::none;
+        return {};
     }
-    profiles[user_count] = user;
+    profiles[user_count] = profile;
     return user_count++;
 }
 
@@ -56,7 +81,7 @@ bool ProfileManager::RemoveProfileAtIndex(std::size_t index) {
 
 /// Helper function to register a user to the system
 ResultCode ProfileManager::AddUser(const ProfileInfo& user) {
-    if (AddToProfiles(user) == boost::none) {
+    if (!AddToProfiles(user)) {
         return ERROR_TOO_MANY_USERS;
     }
     return RESULT_SUCCESS;
@@ -101,31 +126,40 @@ ResultCode ProfileManager::CreateNewUser(UUID uuid, const std::string& username)
     return CreateNewUser(uuid, username_output);
 }
 
+std::optional<UUID> ProfileManager::GetUser(std::size_t index) const {
+    if (index >= MAX_USERS) {
+        return {};
+    }
+
+    return profiles[index].user_uuid;
+}
+
 /// Returns a users profile index based on their user id.
-boost::optional<std::size_t> ProfileManager::GetUserIndex(const UUID& uuid) const {
+std::optional<std::size_t> ProfileManager::GetUserIndex(const UUID& uuid) const {
     if (!uuid) {
-        return boost::none;
+        return {};
     }
-    auto iter = std::find_if(profiles.begin(), profiles.end(),
-                             [&uuid](const ProfileInfo& p) { return p.user_uuid == uuid; });
+
+    const auto iter = std::find_if(profiles.begin(), profiles.end(),
+                                   [&uuid](const ProfileInfo& p) { return p.user_uuid == uuid; });
     if (iter == profiles.end()) {
-        return boost::none;
+        return {};
     }
+
     return static_cast<std::size_t>(std::distance(profiles.begin(), iter));
 }
 
 /// Returns a users profile index based on their profile
-boost::optional<std::size_t> ProfileManager::GetUserIndex(const ProfileInfo& user) const {
+std::optional<std::size_t> ProfileManager::GetUserIndex(const ProfileInfo& user) const {
     return GetUserIndex(user.user_uuid);
 }
 
 /// Returns the data structure used by the switch when GetProfileBase is called on acc:*
-bool ProfileManager::GetProfileBase(boost::optional<std::size_t> index,
-                                    ProfileBase& profile) const {
-    if (index == boost::none || index >= MAX_USERS) {
+bool ProfileManager::GetProfileBase(std::optional<std::size_t> index, ProfileBase& profile) const {
+    if (!index || index >= MAX_USERS) {
         return false;
     }
-    const auto& prof_info = profiles[index.get()];
+    const auto& prof_info = profiles[*index];
     profile.user_uuid = prof_info.user_uuid;
     profile.username = prof_info.username;
     profile.timestamp = prof_info.creation_time;
@@ -134,7 +168,7 @@ bool ProfileManager::GetProfileBase(boost::optional<std::size_t> index,
 
 /// Returns the data structure used by the switch when GetProfileBase is called on acc:*
 bool ProfileManager::GetProfileBase(UUID uuid, ProfileBase& profile) const {
-    auto idx = GetUserIndex(uuid);
+    const auto idx = GetUserIndex(uuid);
     return GetProfileBase(idx, profile);
 }
 
@@ -161,26 +195,34 @@ std::size_t ProfileManager::GetOpenUserCount() const {
 
 /// Checks if a user id exists in our profile manager
 bool ProfileManager::UserExists(UUID uuid) const {
-    return (GetUserIndex(uuid) != boost::none);
+    return GetUserIndex(uuid) != std::nullopt;
+}
+
+bool ProfileManager::UserExistsIndex(std::size_t index) const {
+    if (index >= MAX_USERS)
+        return false;
+    return profiles[index].user_uuid.uuid != INVALID_UUID;
 }
 
 /// Opens a specific user
 void ProfileManager::OpenUser(UUID uuid) {
-    auto idx = GetUserIndex(uuid);
-    if (idx == boost::none) {
+    const auto idx = GetUserIndex(uuid);
+    if (!idx) {
         return;
     }
-    profiles[idx.get()].is_open = true;
+
+    profiles[*idx].is_open = true;
     last_opened_user = uuid;
 }
 
 /// Closes a specific user
 void ProfileManager::CloseUser(UUID uuid) {
-    auto idx = GetUserIndex(uuid);
-    if (idx == boost::none) {
+    const auto idx = GetUserIndex(uuid);
+    if (!idx) {
         return;
     }
-    profiles[idx.get()].is_open = false;
+
+    profiles[*idx].is_open = false;
 }
 
 /// Gets all valid user ids on the system
@@ -210,10 +252,10 @@ UUID ProfileManager::GetLastOpenedUser() const {
 }
 
 /// Return the users profile base and the unknown arbitary data.
-bool ProfileManager::GetProfileBaseAndData(boost::optional<std::size_t> index, ProfileBase& profile,
+bool ProfileManager::GetProfileBaseAndData(std::optional<std::size_t> index, ProfileBase& profile,
                                            ProfileData& data) const {
     if (GetProfileBase(index, profile)) {
-        data = profiles[index.get()].data;
+        data = profiles[*index].data;
         return true;
     }
     return false;
@@ -222,7 +264,7 @@ bool ProfileManager::GetProfileBaseAndData(boost::optional<std::size_t> index, P
 /// Return the users profile base and the unknown arbitary data.
 bool ProfileManager::GetProfileBaseAndData(UUID uuid, ProfileBase& profile,
                                            ProfileData& data) const {
-    auto idx = GetUserIndex(uuid);
+    const auto idx = GetUserIndex(uuid);
     return GetProfileBaseAndData(idx, profile, data);
 }
 
@@ -239,4 +281,96 @@ bool ProfileManager::CanSystemRegisterUser() const {
     // emulate qlaunch. Update this to dynamically change.
 }
 
+bool ProfileManager::RemoveUser(UUID uuid) {
+    const auto index = GetUserIndex(uuid);
+    if (!index) {
+        return false;
+    }
+
+    profiles[*index] = ProfileInfo{};
+    std::stable_partition(profiles.begin(), profiles.end(),
+                          [](const ProfileInfo& profile) { return profile.user_uuid; });
+    return true;
+}
+
+bool ProfileManager::SetProfileBase(UUID uuid, const ProfileBase& profile_new) {
+    const auto index = GetUserIndex(uuid);
+    if (!index || profile_new.user_uuid == UUID(INVALID_UUID)) {
+        return false;
+    }
+
+    auto& profile = profiles[*index];
+    profile.user_uuid = profile_new.user_uuid;
+    profile.username = profile_new.username;
+    profile.creation_time = profile_new.timestamp;
+
+    return true;
+}
+
+void ProfileManager::ParseUserSaveFile() {
+    FileUtil::IOFile save(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                              ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat",
+                          "rb");
+
+    if (!save.IsOpen()) {
+        LOG_WARNING(Service_ACC, "Failed to load profile data from save data... Generating new "
+                                 "user 'yuzu' with random UUID.");
+        return;
+    }
+
+    ProfileDataRaw data;
+    if (save.ReadBytes(&data, sizeof(ProfileDataRaw)) != sizeof(ProfileDataRaw)) {
+        LOG_WARNING(Service_ACC, "profiles.dat is smaller than expected... Generating new user "
+                                 "'yuzu' with random UUID.");
+        return;
+    }
+
+    for (std::size_t i = 0; i < MAX_USERS; ++i) {
+        const auto& user = data.users[i];
+
+        if (user.uuid != UUID(INVALID_UUID))
+            AddUser({user.uuid, user.username, user.timestamp, {}, false});
+    }
+
+    std::stable_partition(profiles.begin(), profiles.end(),
+                          [](const ProfileInfo& profile) { return profile.user_uuid; });
+}
+
+void ProfileManager::WriteUserSaveFile() {
+    ProfileDataRaw raw{};
+
+    for (std::size_t i = 0; i < MAX_USERS; ++i) {
+        raw.users[i].username = profiles[i].username;
+        raw.users[i].uuid2 = profiles[i].user_uuid;
+        raw.users[i].uuid = profiles[i].user_uuid;
+        raw.users[i].timestamp = profiles[i].creation_time;
+    }
+
+    const auto raw_path =
+        FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + "/system/save/8000000000000010";
+    if (FileUtil::Exists(raw_path) && !FileUtil::IsDirectory(raw_path))
+        FileUtil::Delete(raw_path);
+
+    const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
+                      ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat";
+
+    if (!FileUtil::CreateFullPath(path)) {
+        LOG_WARNING(Service_ACC, "Failed to create full path of profiles.dat. Create the directory "
+                                 "nand/system/save/8000000000000010/su/avators to mitigate this "
+                                 "issue.");
+        return;
+    }
+
+    FileUtil::IOFile save(path, "wb");
+
+    if (!save.IsOpen()) {
+        LOG_WARNING(Service_ACC, "Failed to write save data to file... No changes to user data "
+                                 "made in current session will be saved.");
+        return;
+    }
+
+    save.Resize(sizeof(ProfileDataRaw));
+    save.WriteBytes(&raw, sizeof(ProfileDataRaw));
+}
+
 }; // namespace Service::Account

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

@@ -5,8 +5,8 @@
 #pragma once
 
 #include <array>
+#include <optional>
 
-#include "boost/optional.hpp"
 #include "common/common_types.h"
 #include "common/swap.h"
 #include "core/hle/result.h"
@@ -36,7 +36,7 @@ struct UUID {
     }
 
     // TODO(ogniK): Properly generate uuids based on RFC-4122
-    const UUID& Generate();
+    static UUID Generate();
 
     // Set the UUID to {0,0} to be considered an invalid user
     void Invalidate() {
@@ -45,6 +45,15 @@ struct UUID {
     std::string Format() const {
         return fmt::format("0x{:016X}{:016X}", uuid[1], uuid[0]);
     }
+
+    std::string FormatSwitch() const {
+        std::array<u8, 16> s{};
+        std::memcpy(s.data(), uuid.data(), sizeof(u128));
+        return fmt::format("{:02x}{:02x}{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{:02x}-{:02x}{"
+                           ":02x}{:02x}{:02x}{:02x}{:02x}",
+                           s[0], s[1], s[2], s[3], s[4], s[5], s[6], s[7], s[8], s[9], s[10], s[11],
+                           s[12], s[13], s[14], s[15]);
+    }
 };
 static_assert(sizeof(UUID) == 16, "UUID is an invalid size!");
 
@@ -81,18 +90,19 @@ static_assert(sizeof(ProfileBase) == 0x38, "ProfileBase is an invalid size");
 /// objects
 class ProfileManager {
 public:
-    ProfileManager(); // TODO(ogniK): Load from system save
+    ProfileManager();
     ~ProfileManager();
 
     ResultCode AddUser(const ProfileInfo& user);
     ResultCode CreateNewUser(UUID uuid, const ProfileUsername& username);
     ResultCode CreateNewUser(UUID uuid, const std::string& username);
-    boost::optional<std::size_t> GetUserIndex(const UUID& uuid) const;
-    boost::optional<std::size_t> GetUserIndex(const ProfileInfo& user) const;
-    bool GetProfileBase(boost::optional<std::size_t> index, ProfileBase& profile) const;
+    std::optional<UUID> GetUser(std::size_t index) const;
+    std::optional<std::size_t> GetUserIndex(const UUID& uuid) const;
+    std::optional<std::size_t> GetUserIndex(const ProfileInfo& user) const;
+    bool GetProfileBase(std::optional<std::size_t> index, ProfileBase& profile) const;
     bool GetProfileBase(UUID uuid, ProfileBase& profile) const;
     bool GetProfileBase(const ProfileInfo& user, ProfileBase& profile) const;
-    bool GetProfileBaseAndData(boost::optional<std::size_t> index, ProfileBase& profile,
+    bool GetProfileBaseAndData(std::optional<std::size_t> index, ProfileBase& profile,
                                ProfileData& data) const;
     bool GetProfileBaseAndData(UUID uuid, ProfileBase& profile, ProfileData& data) const;
     bool GetProfileBaseAndData(const ProfileInfo& user, ProfileBase& profile,
@@ -100,6 +110,7 @@ public:
     std::size_t GetUserCount() const;
     std::size_t GetOpenUserCount() const;
     bool UserExists(UUID uuid) const;
+    bool UserExistsIndex(std::size_t index) const;
     void OpenUser(UUID uuid);
     void CloseUser(UUID uuid);
     UserIDArray GetOpenUsers() const;
@@ -108,11 +119,17 @@ public:
 
     bool CanSystemRegisterUser() const;
 
+    bool RemoveUser(UUID uuid);
+    bool SetProfileBase(UUID uuid, const ProfileBase& profile_new);
+
 private:
+    void ParseUserSaveFile();
+    void WriteUserSaveFile();
+    std::optional<std::size_t> AddToProfiles(const ProfileInfo& profile);
+    bool RemoveProfileAtIndex(std::size_t index);
+
     std::array<ProfileInfo, MAX_USERS> profiles{};
     std::size_t user_count = 0;
-    boost::optional<std::size_t> AddToProfiles(const ProfileInfo& profile);
-    bool RemoveProfileAtIndex(std::size_t index);
     UUID last_opened_user{INVALID_UUID};
 };
 

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

@@ -4,17 +4,20 @@
 
 #include <array>
 #include <cinttypes>
+#include <cstring>
 #include <stack>
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/kernel/process.h"
+#include "core/hle/service/acc/profile_manager.h"
 #include "core/hle/service/am/am.h"
 #include "core/hle/service/am/applet_ae.h"
 #include "core/hle/service/am/applet_oe.h"
 #include "core/hle/service/am/idle.h"
 #include "core/hle/service/am/omm.h"
 #include "core/hle/service/am/spsm.h"
+#include "core/hle/service/am/tcap.h"
 #include "core/hle/service/apm/apm.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/hle/service/nvflinger/nvflinger.h"
@@ -25,14 +28,29 @@
 
 namespace Service::AM {
 
+constexpr u32 POP_LAUNCH_PARAMETER_MAGIC = 0xC79497CA;
+
+struct LaunchParameters {
+    u32_le magic;
+    u32_le is_account_selected;
+    u128 current_user;
+    INSERT_PADDING_BYTES(0x70);
+};
+static_assert(sizeof(LaunchParameters) == 0x88);
+
 IWindowController::IWindowController() : ServiceFramework("IWindowController") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "CreateWindow"},
         {1, &IWindowController::GetAppletResourceUserId, "GetAppletResourceUserId"},
         {10, &IWindowController::AcquireForegroundRights, "AcquireForegroundRights"},
         {11, nullptr, "ReleaseForegroundRights"},
         {12, nullptr, "RejectToChangeIntoBackground"},
+        {20, nullptr, "SetAppletWindowVisibility"},
+        {21, nullptr, "SetAppletGpuTimeSlice"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
@@ -87,6 +105,7 @@ void IAudioController::GetLibraryAppletExpectedMasterVolume(Kernel::HLERequestCo
 }
 
 IDisplayController::IDisplayController() : ServiceFramework("IDisplayController") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "GetLastForegroundCaptureImage"},
         {1, nullptr, "UpdateLastForegroundCaptureImage"},
@@ -117,7 +136,11 @@ IDisplayController::IDisplayController() : ServiceFramework("IDisplayController"
         {25, nullptr, "ReleaseLastForegroundCaptureSharedBuffer"},
         {26, nullptr, "AcquireCallerAppletCaptureSharedBuffer"},
         {27, nullptr, "ReleaseCallerAppletCaptureSharedBuffer"},
+        // 6.0.0+
+        {28, nullptr, "TakeScreenShotOfOwnLayerEx"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
@@ -128,6 +151,7 @@ IDebugFunctions::~IDebugFunctions() = default;
 
 ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
     : ServiceFramework("ISelfController"), nvflinger(std::move(nvflinger)) {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "Exit"},
         {1, &ISelfController::LockExit, "LockExit"},
@@ -136,10 +160,8 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
         {4, nullptr, "LeaveFatalSection"},
         {9, &ISelfController::GetLibraryAppletLaunchableEvent, "GetLibraryAppletLaunchableEvent"},
         {10, &ISelfController::SetScreenShotPermission, "SetScreenShotPermission"},
-        {11, &ISelfController::SetOperationModeChangedNotification,
-         "SetOperationModeChangedNotification"},
-        {12, &ISelfController::SetPerformanceModeChangedNotification,
-         "SetPerformanceModeChangedNotification"},
+        {11, &ISelfController::SetOperationModeChangedNotification, "SetOperationModeChangedNotification"},
+        {12, &ISelfController::SetPerformanceModeChangedNotification, "SetPerformanceModeChangedNotification"},
         {13, &ISelfController::SetFocusHandlingMode, "SetFocusHandlingMode"},
         {14, &ISelfController::SetRestartMessageEnabled, "SetRestartMessageEnabled"},
         {15, nullptr, "SetScreenShotAppletIdentityInfo"},
@@ -165,7 +187,12 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
         {69, nullptr, "IsAutoSleepDisabled"},
         {70, nullptr, "ReportMultimediaError"},
         {80, nullptr, "SetWirelessPriorityMode"},
+        {90, nullptr, "GetAccumulatedSuspendedTickValue"},
+        {91, nullptr, "GetAccumulatedSuspendedTickChangedEvent"},
+        {1000, nullptr, "GetDebugStorageChannel"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 
     auto& kernel = Core::System::GetInstance().Kernel();
@@ -312,6 +339,7 @@ void ISelfController::GetIdleTimeDetectionExtension(Kernel::HLERequestContext& c
 }
 
 ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, &ICommonStateGetter::GetEventHandle, "GetEventHandle"},
         {1, &ICommonStateGetter::ReceiveMessage, "ReceiveMessage"},
@@ -336,11 +364,12 @@ ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter"
         {52, nullptr, "SwitchLcdBacklight"},
         {55, nullptr, "IsInControllerFirmwareUpdateSection"},
         {60, &ICommonStateGetter::GetDefaultDisplayResolution, "GetDefaultDisplayResolution"},
-        {61, &ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent,
-         "GetDefaultDisplayResolutionChangeEvent"},
+        {61, &ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent, "GetDefaultDisplayResolutionChangeEvent"},
         {62, nullptr, "GetHdcpAuthenticationState"},
         {63, nullptr, "GetHdcpAuthenticationStateChangeEvent"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 
     auto& kernel = Core::System::GetInstance().Kernel();
@@ -432,11 +461,14 @@ class IStorageAccessor final : public ServiceFramework<IStorageAccessor> {
 public:
     explicit IStorageAccessor(std::vector<u8> buffer)
         : ServiceFramework("IStorageAccessor"), buffer(std::move(buffer)) {
+        // clang-format off
         static const FunctionInfo functions[] = {
             {0, &IStorageAccessor::GetSize, "GetSize"},
             {10, &IStorageAccessor::Write, "Write"},
             {11, &IStorageAccessor::Read, "Read"},
         };
+        // clang-format on
+
         RegisterHandlers(functions);
     }
 
@@ -489,10 +521,13 @@ class IStorage final : public ServiceFramework<IStorage> {
 public:
     explicit IStorage(std::vector<u8> buffer)
         : ServiceFramework("IStorage"), buffer(std::move(buffer)) {
+        // clang-format off
         static const FunctionInfo functions[] = {
             {0, &IStorage::Open, "Open"},
             {1, nullptr, "OpenTransferStorage"},
         };
+        // clang-format on
+
         RegisterHandlers(functions);
     }
 
@@ -512,6 +547,7 @@ private:
 class ILibraryAppletAccessor final : public ServiceFramework<ILibraryAppletAccessor> {
 public:
     explicit ILibraryAppletAccessor() : ServiceFramework("ILibraryAppletAccessor") {
+        // clang-format off
         static const FunctionInfo functions[] = {
             {0, &ILibraryAppletAccessor::GetAppletStateChangedEvent, "GetAppletStateChangedEvent"},
             {1, nullptr, "IsCompleted"},
@@ -532,6 +568,8 @@ public:
             {150, nullptr, "RequestForAppletToGetForeground"},
             {160, nullptr, "GetIndirectLayerConsumerHandle"},
         };
+        // clang-format on
+
         RegisterHandlers(functions);
 
         auto& kernel = Core::System::GetInstance().Kernel();
@@ -624,13 +662,13 @@ void ILibraryAppletCreator::CreateStorage(Kernel::HLERequestContext& ctx) {
 }
 
 IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationFunctions") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {1, &IApplicationFunctions::PopLaunchParameter, "PopLaunchParameter"},
         {10, nullptr, "CreateApplicationAndPushAndRequestToStart"},
         {11, nullptr, "CreateApplicationAndPushAndRequestToStartForQuest"},
         {12, nullptr, "CreateApplicationAndRequestToStart"},
-        {13, &IApplicationFunctions::CreateApplicationAndRequestToStartForQuest,
-         "CreateApplicationAndRequestToStartForQuest"},
+        {13, &IApplicationFunctions::CreateApplicationAndRequestToStartForQuest, "CreateApplicationAndRequestToStartForQuest"},
         {20, &IApplicationFunctions::EnsureSaveData, "EnsureSaveData"},
         {21, &IApplicationFunctions::GetDesiredLanguage, "GetDesiredLanguage"},
         {22, &IApplicationFunctions::SetTerminateResult, "SetTerminateResult"},
@@ -638,10 +676,8 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
         {24, nullptr, "GetLaunchStorageInfoForDebug"},
         {25, nullptr, "ExtendSaveData"},
         {26, nullptr, "GetSaveDataSize"},
-        {30, &IApplicationFunctions::BeginBlockingHomeButtonShortAndLongPressed,
-         "BeginBlockingHomeButtonShortAndLongPressed"},
-        {31, &IApplicationFunctions::EndBlockingHomeButtonShortAndLongPressed,
-         "EndBlockingHomeButtonShortAndLongPressed"},
+        {30, &IApplicationFunctions::BeginBlockingHomeButtonShortAndLongPressed, "BeginBlockingHomeButtonShortAndLongPressed"},
+        {31, &IApplicationFunctions::EndBlockingHomeButtonShortAndLongPressed, "EndBlockingHomeButtonShortAndLongPressed"},
         {32, &IApplicationFunctions::BeginBlockingHomeButton, "BeginBlockingHomeButton"},
         {33, &IApplicationFunctions::EndBlockingHomeButton, "EndBlockingHomeButton"},
         {40, &IApplicationFunctions::NotifyRunning, "NotifyRunning"},
@@ -666,6 +702,8 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
         {1000, nullptr, "CreateMovieMaker"},
         {1001, nullptr, "PrepareForJit"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
@@ -698,20 +736,23 @@ void IApplicationFunctions::EndBlockingHomeButton(Kernel::HLERequestContext& ctx
 }
 
 void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) {
-    constexpr std::array<u8, 0x88> data{{
-        0xca, 0x97, 0x94, 0xc7, // Magic
-        1,    0,    0,    0,    // IsAccountSelected (bool)
-        1,    0,    0,    0,    // User Id (word 0)
-        0,    0,    0,    0,    // User Id (word 1)
-        0,    0,    0,    0,    // User Id (word 2)
-        0,    0,    0,    0     // User Id (word 3)
-    }};
+    LaunchParameters params{};
 
-    std::vector<u8> buffer(data.begin(), data.end());
+    params.magic = POP_LAUNCH_PARAMETER_MAGIC;
+    params.is_account_selected = 1;
+
+    Account::ProfileManager profile_manager{};
+    const auto uuid = profile_manager.GetUser(Settings::values.current_user);
+    ASSERT(uuid != std::nullopt);
+    params.current_user = uuid->uuid;
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
 
     rb.Push(RESULT_SUCCESS);
+
+    std::vector<u8> buffer(sizeof(LaunchParameters));
+    std::memcpy(buffer.data(), &params, buffer.size());
+
     rb.PushIpcInterface<AM::IStorage>(buffer);
 
     LOG_DEBUG(Service_AM, "called");
@@ -804,9 +845,11 @@ void InstallInterfaces(SM::ServiceManager& service_manager,
     std::make_shared<IdleSys>()->InstallAsService(service_manager);
     std::make_shared<OMM>()->InstallAsService(service_manager);
     std::make_shared<SPSM>()->InstallAsService(service_manager);
+    std::make_shared<TCAP>()->InstallAsService(service_manager);
 }
 
 IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {10, &IHomeMenuFunctions::RequestToGetForeground, "RequestToGetForeground"},
         {11, nullptr, "LockForeground"},
@@ -815,7 +858,10 @@ IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions"
         {21, nullptr, "GetPopFromGeneralChannelEvent"},
         {30, nullptr, "GetHomeButtonWriterLockAccessor"},
         {31, nullptr, "GetWriterLockAccessorEx"},
+        {100, nullptr, "PopRequestLaunchApplicationForDebug"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
@@ -828,6 +874,7 @@ void IHomeMenuFunctions::RequestToGetForeground(Kernel::HLERequestContext& ctx)
 }
 
 IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStateController") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "RequestToEnterSleep"},
         {1, nullptr, "EnterSleep"},
@@ -841,18 +888,23 @@ IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStat
         {14, nullptr, "ShouldSleepOnBoot"},
         {15, nullptr, "GetHdcpAuthenticationFailedEvent"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
 IGlobalStateController::~IGlobalStateController() = default;
 
 IApplicationCreator::IApplicationCreator() : ServiceFramework("IApplicationCreator") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "CreateApplication"},
         {1, nullptr, "PopLaunchRequestedApplication"},
         {10, nullptr, "CreateSystemApplication"},
         {100, nullptr, "PopFloatingApplicationForDevelopment"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 
@@ -860,6 +912,7 @@ IApplicationCreator::~IApplicationCreator() = default;
 
 IProcessWindingController::IProcessWindingController()
     : ServiceFramework("IProcessWindingController") {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {0, nullptr, "GetLaunchReason"},
         {11, nullptr, "OpenCallingLibraryApplet"},
@@ -870,6 +923,8 @@ IProcessWindingController::IProcessWindingController()
         {40, nullptr, "ReserveToStartAndWaitAndUnwindThis"},
         {41, nullptr, "ReserveToStartAndWait"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 

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

@@ -211,6 +211,7 @@ void AppletAE::OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx) {
 
 AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
     : ServiceFramework("appletAE"), nvflinger(std::move(nvflinger)) {
+    // clang-format off
     static const FunctionInfo functions[] = {
         {100, &AppletAE::OpenSystemAppletProxy, "OpenSystemAppletProxy"},
         {200, &AppletAE::OpenLibraryAppletProxyOld, "OpenLibraryAppletProxyOld"},
@@ -218,7 +219,10 @@ AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
         {300, nullptr, "OpenOverlayAppletProxy"},
         {350, nullptr, "OpenSystemApplicationProxy"},
         {400, nullptr, "CreateSelfLibraryAppletCreatorForDevelop"},
+        {401, nullptr, "GetSystemAppletControllerForDebug"},
     };
+    // clang-format on
+
     RegisterHandlers(functions);
 }
 

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

@@ -14,6 +14,7 @@ class IApplicationProxy final : public ServiceFramework<IApplicationProxy> {
 public:
     explicit IApplicationProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
         : ServiceFramework("IApplicationProxy"), nvflinger(std::move(nvflinger)) {
+        // clang-format off
         static const FunctionInfo functions[] = {
             {0, &IApplicationProxy::GetCommonStateGetter, "GetCommonStateGetter"},
             {1, &IApplicationProxy::GetSelfController, "GetSelfController"},
@@ -25,6 +26,8 @@ public:
             {20, &IApplicationProxy::GetApplicationFunctions, "GetApplicationFunctions"},
             {1000, &IApplicationProxy::GetDebugFunctions, "GetDebugFunctions"},
         };
+        // clang-format on
+
         RegisterHandlers(functions);
     }
 

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

@@ -12,9 +12,9 @@ IdleSys::IdleSys() : ServiceFramework{"idle:sys"} {
         {0, nullptr, "GetAutoPowerDownEvent"},
         {1, nullptr, "Unknown1"},
         {2, nullptr, "Unknown2"},
-        {3, nullptr, "Unknown3"},
-        {4, nullptr, "Unknown4"},
-        {5, nullptr, "Unknown5"},
+        {3, nullptr, "SetHandlingContext"},
+        {4, nullptr, "LoadAndApplySettings"},
+        {5, nullptr, "ReportUserIsActive"},
     };
     // clang-format on
 

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

@@ -0,0 +1,23 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/hle/service/am/tcap.h"
+
+namespace Service::AM {
+
+TCAP::TCAP() : ServiceFramework{"tcap"} {
+    // clang-format off
+    static const FunctionInfo functions[] = {
+        {0, nullptr, "GetContinuousHighSkinTemperatureEvent"},
+        {1, nullptr, "SetOperationMode"},
+        {2, nullptr, "LoadAndApplySettings"},
+    };
+    // clang-format on
+
+    RegisterHandlers(functions);
+}
+
+TCAP::~TCAP() = default;
+
+} // namespace Service::AM

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

@@ -0,0 +1,17 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/service/service.h"
+
+namespace Service::AM {
+
+class TCAP final : public ServiceFramework<TCAP> {
+public:
+    explicit TCAP();
+    ~TCAP() override;
+};
+
+} // namespace Service::AM

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

@@ -24,8 +24,8 @@ namespace Service::AOC {
 constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000;
 constexpr u64 DLC_BASE_TO_AOC_ID = 0x1000;
 
-static bool CheckAOCTitleIDMatchesBase(u64 base, u64 aoc) {
-    return (aoc & DLC_BASE_TITLE_ID_MASK) == base;
+static bool CheckAOCTitleIDMatchesBase(u64 title_id, u64 base) {
+    return (title_id & DLC_BASE_TITLE_ID_MASK) == base;
 }
 
 static std::vector<u64> AccumulateAOCTitleIDs() {
@@ -74,7 +74,7 @@ void AOC_U::CountAddOnContent(Kernel::HLERequestContext& ctx) {
     const auto current = Core::System::GetInstance().CurrentProcess()->GetTitleID();
     rb.Push<u32>(static_cast<u32>(
         std::count_if(add_on_content.begin(), add_on_content.end(),
-                      [&current](u64 tid) { return (tid & DLC_BASE_TITLE_ID_MASK) == current; })));
+                      [current](u64 tid) { return CheckAOCTitleIDMatchesBase(tid, current); })));
 }
 
 void AOC_U::ListAddOnContent(Kernel::HLERequestContext& ctx) {

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

@@ -17,6 +17,7 @@
 #include "core/file_sys/errors.h"
 #include "core/file_sys/mode.h"
 #include "core/file_sys/nca_metadata.h"
+#include "core/file_sys/patch_manager.h"
 #include "core/file_sys/savedata_factory.h"
 #include "core/file_sys/vfs.h"
 #include "core/hle/ipc_helpers.h"
@@ -630,6 +631,7 @@ void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) {
               static_cast<u8>(storage_id), unknown, title_id);
 
     auto data = OpenRomFS(title_id, storage_id, FileSys::ContentRecordType::Data);
+
     if (data.Failed()) {
         // TODO(DarkLordZach): Find the right error code to use here
         LOG_ERROR(Service_FS,
@@ -640,7 +642,9 @@ void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) {
         return;
     }
 
-    IStorage storage(std::move(data.Unwrap()));
+    FileSys::PatchManager pm{title_id};
+
+    IStorage storage(pm.PatchRomFS(std::move(data.Unwrap()), 0, FileSys::ContentRecordType::Data));
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);

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

@@ -26,7 +26,11 @@ constexpr s32 HID_JOYSTICK_MAX = 0x7fff;
 constexpr s32 HID_JOYSTICK_MIN = -0x7fff;
 constexpr std::size_t NPAD_OFFSET = 0x9A00;
 constexpr u32 BATTERY_FULL = 2;
-
+constexpr u32 NPAD_HANDHELD = 32;
+constexpr u32 NPAD_UNKNOWN = 16; // TODO(ogniK): What is this?
+constexpr u32 MAX_NPAD_ID = 7;
+constexpr Controller_NPad::NPadControllerType PREFERRED_CONTROLLER =
+    Controller_NPad::NPadControllerType::JoyDual;
 constexpr std::array<u32, 10> npad_id_list{
     0, 1, 2, 3, 4, 5, 6, 7, 32, 16,
 };
@@ -104,9 +108,10 @@ void Controller_NPad::OnInit() {
     styleset_changed_event =
         Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "npad:NpadStyleSetChanged");
 
-    if (!IsControllerActivated())
+    if (!IsControllerActivated()) {
         return;
-    std::size_t controller{};
+    }
+
     if (style.raw == 0) {
         // We want to support all controllers
         style.handheld.Assign(1);
@@ -121,7 +126,7 @@ void Controller_NPad::OnInit() {
         supported_npad_id_types.resize(npad_id_list.size());
         std::memcpy(supported_npad_id_types.data(), npad_id_list.data(),
                     npad_id_list.size() * sizeof(u32));
-        AddNewController(NPadControllerType::JoyDual);
+        AddNewController(PREFERRED_CONTROLLER);
     }
 }
 
@@ -218,6 +223,51 @@ void Controller_NPad::OnUpdate(u8* data, std::size_t data_len) {
         rstick_entry.x = static_cast<s32>(stick_r_x_f * HID_JOYSTICK_MAX);
         rstick_entry.y = static_cast<s32>(stick_r_y_f * HID_JOYSTICK_MAX);
 
+        if (controller_type == NPadControllerType::JoyLeft ||
+            controller_type == NPadControllerType::JoyRight) {
+            if (npad.properties.is_horizontal) {
+                ControllerPadState state{};
+                AnalogPosition temp_lstick_entry{};
+                AnalogPosition temp_rstick_entry{};
+                if (controller_type == NPadControllerType::JoyLeft) {
+                    state.d_down.Assign(pad_state.d_left.Value());
+                    state.d_left.Assign(pad_state.d_up.Value());
+                    state.d_right.Assign(pad_state.d_down.Value());
+                    state.d_up.Assign(pad_state.d_right.Value());
+                    state.l.Assign(pad_state.l.Value() | pad_state.sl.Value());
+                    state.r.Assign(pad_state.r.Value() | pad_state.sr.Value());
+
+                    state.zl.Assign(pad_state.zl.Value());
+                    state.plus.Assign(pad_state.minus.Value());
+
+                    temp_lstick_entry = lstick_entry;
+                    temp_rstick_entry = rstick_entry;
+                    std::swap(temp_lstick_entry.x, temp_lstick_entry.y);
+                    std::swap(temp_rstick_entry.x, temp_rstick_entry.y);
+                    temp_lstick_entry.y *= -1;
+                } else if (controller_type == NPadControllerType::JoyRight) {
+                    state.x.Assign(pad_state.a.Value());
+                    state.a.Assign(pad_state.b.Value());
+                    state.b.Assign(pad_state.y.Value());
+                    state.y.Assign(pad_state.b.Value());
+
+                    state.l.Assign(pad_state.l.Value() | pad_state.sl.Value());
+                    state.r.Assign(pad_state.r.Value() | pad_state.sr.Value());
+                    state.zr.Assign(pad_state.zr.Value());
+                    state.plus.Assign(pad_state.plus.Value());
+
+                    temp_lstick_entry = lstick_entry;
+                    temp_rstick_entry = rstick_entry;
+                    std::swap(temp_lstick_entry.x, temp_lstick_entry.y);
+                    std::swap(temp_rstick_entry.x, temp_rstick_entry.y);
+                    temp_rstick_entry.x *= -1;
+                }
+                pad_state.raw = state.raw;
+                lstick_entry = temp_lstick_entry;
+                rstick_entry = temp_rstick_entry;
+            }
+        }
+
         auto& main_controller =
             npad.main_controller_states.npad[npad.main_controller_states.common.last_entry_index];
         auto& handheld_entry =
@@ -320,6 +370,16 @@ void Controller_NPad::SetSupportedNPadIdTypes(u8* data, std::size_t length) {
     supported_npad_id_types.clear();
     supported_npad_id_types.resize(length / sizeof(u32));
     std::memcpy(supported_npad_id_types.data(), data, length);
+    for (std::size_t i = 0; i < connected_controllers.size(); i++) {
+        auto& controller = connected_controllers[i];
+        if (!controller.is_connected) {
+            continue;
+        }
+        if (!IsControllerSupported(PREFERRED_CONTROLLER)) {
+            controller.type = DecideBestController(PREFERRED_CONTROLLER);
+            InitNewlyAddedControler(i);
+        }
+    }
 }
 
 void Controller_NPad::GetSupportedNpadIdTypes(u32* data, std::size_t max_length) {
@@ -351,11 +411,11 @@ void Controller_NPad::VibrateController(const std::vector<u32>& controller_ids,
     for (std::size_t i = 0; i < controller_ids.size(); i++) {
         std::size_t controller_pos = i;
         // Handheld controller conversion
-        if (controller_pos == 32) {
+        if (controller_pos == NPAD_HANDHELD) {
             controller_pos = 8;
         }
         // Unknown controller conversion
-        if (controller_pos == 16) {
+        if (controller_pos == NPAD_UNKNOWN) {
             controller_pos = 9;
         }
         if (connected_controllers[controller_pos].is_connected) {
@@ -433,4 +493,128 @@ Controller_NPad::LedPattern Controller_NPad::GetLedPattern(u32 npad_id) {
 void Controller_NPad::SetVibrationEnabled(bool can_vibrate) {
     can_controllers_vibrate = can_vibrate;
 }
+
+bool Controller_NPad::IsControllerSupported(NPadControllerType controller) const {
+    const bool support_handheld =
+        std::find(supported_npad_id_types.begin(), supported_npad_id_types.end(), NPAD_HANDHELD) !=
+        supported_npad_id_types.end();
+    if (controller == NPadControllerType::Handheld) {
+        // Handheld is not even a supported type, lets stop here
+        if (!support_handheld) {
+            return false;
+        }
+        // Handheld should not be supported in docked mode
+        if (Settings::values.use_docked_mode) {
+            return false;
+        }
+
+        return true;
+    }
+    if (std::any_of(supported_npad_id_types.begin(), supported_npad_id_types.end(),
+                    [](u32 npad_id) { return npad_id <= MAX_NPAD_ID; })) {
+        switch (controller) {
+        case NPadControllerType::ProController:
+            return style.pro_controller;
+        case NPadControllerType::JoyDual:
+            return style.joycon_dual;
+        case NPadControllerType::JoyLeft:
+            return style.joycon_left;
+        case NPadControllerType::JoyRight:
+            return style.joycon_right;
+        case NPadControllerType::Pokeball:
+            return style.pokeball;
+        default:
+            return false;
+        }
+    }
+    return false;
+}
+
+Controller_NPad::NPadControllerType Controller_NPad::DecideBestController(
+    NPadControllerType priority) const {
+    if (IsControllerSupported(priority)) {
+        return priority;
+    }
+    const auto is_docked = Settings::values.use_docked_mode;
+    if (is_docked && priority == NPadControllerType::Handheld) {
+        priority = NPadControllerType::JoyDual;
+        if (IsControllerSupported(priority)) {
+            return priority;
+        }
+    }
+    std::vector<NPadControllerType> priority_list;
+    switch (priority) {
+    case NPadControllerType::ProController:
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::Handheld:
+        priority_list.push_back(NPadControllerType::JoyDual);
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::JoyDual:
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::JoyLeft:
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::JoyRight:
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::Pokeball);
+        break;
+    case NPadControllerType::Pokeball:
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        break;
+    default:
+        priority_list.push_back(NPadControllerType::JoyDual);
+        if (!is_docked) {
+            priority_list.push_back(NPadControllerType::Handheld);
+        }
+        priority_list.push_back(NPadControllerType::ProController);
+        priority_list.push_back(NPadControllerType::JoyLeft);
+        priority_list.push_back(NPadControllerType::JoyRight);
+        priority_list.push_back(NPadControllerType::JoyDual);
+    }
+
+    const auto iter = std::find_if(priority_list.begin(), priority_list.end(),
+                                   [this](auto type) { return IsControllerSupported(type); });
+    if (iter == priority_list.end()) {
+        UNIMPLEMENTED_MSG("Could not find supported controller!");
+        return priority;
+    }
+
+    return *iter;
+}
+
 } // namespace Service::HID

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

@@ -283,5 +283,7 @@ private:
     bool can_controllers_vibrate{true};
 
     void InitNewlyAddedControler(std::size_t controller_idx);
+    bool IsControllerSupported(NPadControllerType controller) const;
+    NPadControllerType DecideBestController(NPadControllerType priority) const;
 };
 } // namespace Service::HID

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

@@ -827,6 +827,7 @@ public:
             {11, nullptr, "EnableJoyPollingReceiveMode"},
             {12, nullptr, "DisableJoyPollingReceiveMode"},
             {13, nullptr, "GetPollingData"},
+            {14, nullptr, "SetStatusManagerType"},
         };
         // clang-format on
 

src/core/hle/service/lbl/lbl.cpp

@@ -18,35 +18,35 @@ public:
     explicit LBL() : ServiceFramework{"lbl"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "Unknown1"},
-            {1, nullptr, "Unknown2"},
-            {2, nullptr, "Unknown3"},
-            {3, nullptr, "GetCurrentBacklightLevel"},
-            {4, nullptr, "Unknown4"},
-            {5, nullptr, "GetAlsComputedBacklightLevel"},
-            {6, nullptr, "TurnOffBacklight"},
-            {7, nullptr, "TurnOnBacklight"},
-            {8, nullptr, "GetBacklightStatus"},
-            {9, nullptr, "Unknown5"},
-            {10, nullptr, "Unknown6"},
-            {11, nullptr, "Unknown7"},
-            {12, nullptr, "Unknown8"},
-            {13, nullptr, "Unknown9"},
-            {14, nullptr, "Unknown10"},
-            {15, nullptr, "GetAutoBrightnessSetting"},
-            {16, nullptr, "ReadRawLightSensor"},
-            {17, nullptr, "Unknown11"},
-            {18, nullptr, "Unknown12"},
-            {19, nullptr, "Unknown13"},
-            {20, nullptr, "Unknown14"},
-            {21, nullptr, "Unknown15"},
-            {22, nullptr, "Unknown16"},
-            {23, nullptr, "Unknown17"},
-            {24, nullptr, "Unknown18"},
-            {25, nullptr, "Unknown19"},
+            {0, nullptr, "SaveCurrentSetting"},
+            {1, nullptr, "LoadCurrentSetting"},
+            {2, nullptr, "SetCurrentBrightnessSetting"},
+            {3, nullptr, "GetCurrentBrightnessSetting"},
+            {4, nullptr, "ApplyCurrentBrightnessSettingToBacklight"},
+            {5, nullptr, "GetBrightnessSettingAppliedToBacklight"},
+            {6, nullptr, "SwitchBacklightOn"},
+            {7, nullptr, "SwitchBacklightOff"},
+            {8, nullptr, "GetBacklightSwitchStatus"},
+            {9, nullptr, "EnableDimming"},
+            {10, nullptr, "DisableDimming"},
+            {11, nullptr, "IsDimmingEnabled"},
+            {12, nullptr, "EnableAutoBrightnessControl"},
+            {13, nullptr, "DisableAutoBrightnessControl"},
+            {14, nullptr, "IsAutoBrightnessControlEnabled"},
+            {15, nullptr, "SetAmbientLightSensorValue"},
+            {16, nullptr, "GetAmbientLightSensorValue"},
+            {17, nullptr, "SetBrightnessReflectionDelayLevel"},
+            {18, nullptr, "GetBrightnessReflectionDelayLevel"},
+            {19, nullptr, "SetCurrentBrightnessMapping"},
+            {20, nullptr, "GetCurrentBrightnessMapping"},
+            {21, nullptr, "SetCurrentAmbientLightSensorMapping"},
+            {22, nullptr, "GetCurrentAmbientLightSensorMapping"},
+            {23, nullptr, "IsAmbientLightSensorAvailable"},
+            {24, nullptr, "SetCurrentBrightnessSettingForVrMode"},
+            {25, nullptr, "GetCurrentBrightnessSettingForVrMode"},
             {26, &LBL::EnableVrMode, "EnableVrMode"},
             {27, &LBL::DisableVrMode, "DisableVrMode"},
-            {28, &LBL::GetVrMode, "GetVrMode"},
+            {28, &LBL::IsVrModeEnabled, "IsVrModeEnabled"},
         };
         // clang-format on
 
@@ -72,7 +72,7 @@ private:
         LOG_DEBUG(Service_LBL, "called");
     }
 
-    void GetVrMode(Kernel::HLERequestContext& ctx) {
+    void IsVrModeEnabled(Kernel::HLERequestContext& ctx) {
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
         rb.Push(vr_mode_enabled);

src/core/hle/service/ldr/ldr.cpp

@@ -3,9 +3,13 @@
 // Refer to the license.txt file included.
 
 #include <memory>
+#include <fmt/format.h>
 
+#include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/process.h"
 #include "core/hle/service/ldr/ldr.h"
 #include "core/hle/service/service.h"
+#include "core/loader/nro.h"
 
 namespace Service::LDR {
 
@@ -59,16 +63,58 @@ public:
     explicit RelocatableObject() : ServiceFramework{"ldr:ro"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "LoadNro"},
+            {0, &RelocatableObject::LoadNro, "LoadNro"},
             {1, nullptr, "UnloadNro"},
-            {2, nullptr, "LoadNrr"},
+            {2, &RelocatableObject::LoadNrr, "LoadNrr"},
             {3, nullptr, "UnloadNrr"},
-            {4, nullptr, "Initialize"},
+            {4, &RelocatableObject::Initialize, "Initialize"},
         };
         // clang-format on
 
         RegisterHandlers(functions);
     }
+
+    void LoadNrr(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_WARNING(Service_LDR, "(STUBBED) called");
+    }
+
+    void LoadNro(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        rp.Skip(2, false);
+        const VAddr nro_addr{rp.Pop<VAddr>()};
+        const u64 nro_size{rp.Pop<u64>()};
+        const VAddr bss_addr{rp.Pop<VAddr>()};
+        const u64 bss_size{rp.Pop<u64>()};
+
+        // Read NRO data from memory
+        std::vector<u8> nro_data(nro_size);
+        Memory::ReadBlock(nro_addr, nro_data.data(), nro_size);
+
+        // Load NRO as new executable module
+        const VAddr addr{*Core::CurrentProcess()->VMManager().FindFreeRegion(nro_size + bss_size)};
+        Loader::AppLoader_NRO::LoadNro(nro_data, fmt::format("nro-{:08x}", addr), addr);
+
+        // TODO(bunnei): This is an incomplete implementation. It was tested with Super Mario Party.
+        // It is currently missing:
+        // - Signature checks with LoadNRR
+        // - Checking if a module has already been loaded
+        // - Using/validating BSS, etc. params (these are used from NRO header instead)
+        // - Error checking
+        // - ...Probably other things
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push(addr);
+        LOG_WARNING(Service_LDR, "(STUBBED) called");
+    }
+
+    void Initialize(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_WARNING(Service_LDR, "(STUBBED) called");
+    }
 };
 
 void InstallInterfaces(SM::ServiceManager& sm) {

src/core/hle/service/nfc/nfc.cpp

@@ -10,12 +10,13 @@
 #include "core/hle/service/nfc/nfc.h"
 #include "core/hle/service/service.h"
 #include "core/hle/service/sm/sm.h"
+#include "core/settings.h"
 
 namespace Service::NFC {
 
 class IAm final : public ServiceFramework<IAm> {
 public:
-    explicit IAm() : ServiceFramework{"IAm"} {
+    explicit IAm() : ServiceFramework{"NFC::IAm"} {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "Initialize"},
@@ -52,7 +53,7 @@ private:
 
 class MFIUser final : public ServiceFramework<MFIUser> {
 public:
-    explicit MFIUser() : ServiceFramework{"IUser"} {
+    explicit MFIUser() : ServiceFramework{"NFC::MFIUser"} {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "Initialize"},
@@ -100,13 +101,13 @@ private:
 
 class IUser final : public ServiceFramework<IUser> {
 public:
-    explicit IUser() : ServiceFramework{"IUser"} {
+    explicit IUser() : ServiceFramework{"NFC::IUser"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "Initialize"},
-            {1, nullptr, "Finalize"},
-            {2, nullptr, "GetState"},
-            {3, nullptr, "IsNfcEnabled"},
+            {0, &IUser::InitializeOld, "InitializeOld"},
+            {1, &IUser::FinalizeOld, "FinalizeOld"},
+            {2, &IUser::GetStateOld, "GetStateOld"},
+            {3, &IUser::IsNfcEnabledOld, "IsNfcEnabledOld"},
             {400, nullptr, "Initialize"},
             {401, nullptr, "Finalize"},
             {402, nullptr, "GetState"},
@@ -130,11 +131,47 @@ public:
 
         RegisterHandlers(functions);
     }
+
+private:
+    enum class NfcStates : u32 {
+        Finalized = 6,
+    };
+
+    void InitializeOld(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0};
+        rb.Push(RESULT_SUCCESS);
+
+        // We don't deal with hardware initialization so we can just stub this.
+        LOG_DEBUG(Service_NFC, "called");
+    }
+
+    void IsNfcEnabledOld(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u8>(Settings::values.enable_nfc);
+
+        LOG_DEBUG(Service_NFC, "IsNfcEnabledOld");
+    }
+
+    void GetStateOld(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFC, "(STUBBED) called");
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushEnum(NfcStates::Finalized); // TODO(ogniK): Figure out if this matches nfp
+    }
+
+    void FinalizeOld(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFC, "(STUBBED) called");
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
 };
 
 class NFC_U final : public ServiceFramework<NFC_U> {
 public:
-    explicit NFC_U() : ServiceFramework{"nfc:u"} {
+    explicit NFC_U() : ServiceFramework{"nfc:user"} {
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, &NFC_U::CreateUserInterface, "CreateUserInterface"},

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

@@ -2,56 +2,67 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <atomic>
+
 #include "common/logging/log.h"
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
+#include "core/hle/lock.h"
 #include "core/hle/service/hid/hid.h"
 #include "core/hle/service/nfp/nfp.h"
 #include "core/hle/service/nfp/nfp_user.h"
 
 namespace Service::NFP {
 
+namespace ErrCodes {
+constexpr ResultCode ERR_TAG_FAILED(ErrorModule::NFP,
+                                    -1); // TODO(ogniK): Find the actual error code
+}
+
 Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
-    : ServiceFramework(name), module(std::move(module)) {}
+    : ServiceFramework(name), module(std::move(module)) {
+    auto& kernel = Core::System::GetInstance().Kernel();
+    nfc_tag_load =
+        Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "IUser:NFCTagDetected");
+}
 
 Module::Interface::~Interface() = default;
 
 class IUser final : public ServiceFramework<IUser> {
 public:
-    IUser() : ServiceFramework("IUser") {
+    IUser(Module::Interface& nfp_interface)
+        : ServiceFramework("NFP::IUser"), nfp_interface(nfp_interface) {
         static const FunctionInfo functions[] = {
             {0, &IUser::Initialize, "Initialize"},
-            {1, nullptr, "Finalize"},
+            {1, &IUser::Finalize, "Finalize"},
             {2, &IUser::ListDevices, "ListDevices"},
-            {3, nullptr, "StartDetection"},
-            {4, nullptr, "StopDetection"},
-            {5, nullptr, "Mount"},
-            {6, nullptr, "Unmount"},
-            {7, nullptr, "OpenApplicationArea"},
-            {8, nullptr, "GetApplicationArea"},
+            {3, &IUser::StartDetection, "StartDetection"},
+            {4, &IUser::StopDetection, "StopDetection"},
+            {5, &IUser::Mount, "Mount"},
+            {6, &IUser::Unmount, "Unmount"},
+            {7, &IUser::OpenApplicationArea, "OpenApplicationArea"},
+            {8, &IUser::GetApplicationArea, "GetApplicationArea"},
             {9, nullptr, "SetApplicationArea"},
             {10, nullptr, "Flush"},
             {11, nullptr, "Restore"},
             {12, nullptr, "CreateApplicationArea"},
-            {13, nullptr, "GetTagInfo"},
-            {14, nullptr, "GetRegisterInfo"},
-            {15, nullptr, "GetCommonInfo"},
-            {16, nullptr, "GetModelInfo"},
+            {13, &IUser::GetTagInfo, "GetTagInfo"},
+            {14, &IUser::GetRegisterInfo, "GetRegisterInfo"},
+            {15, &IUser::GetCommonInfo, "GetCommonInfo"},
+            {16, &IUser::GetModelInfo, "GetModelInfo"},
             {17, &IUser::AttachActivateEvent, "AttachActivateEvent"},
             {18, &IUser::AttachDeactivateEvent, "AttachDeactivateEvent"},
             {19, &IUser::GetState, "GetState"},
             {20, &IUser::GetDeviceState, "GetDeviceState"},
             {21, &IUser::GetNpadId, "GetNpadId"},
-            {22, nullptr, "GetApplicationArea2"},
+            {22, &IUser::GetApplicationAreaSize, "GetApplicationAreaSize"},
             {23, &IUser::AttachAvailabilityChangeEvent, "AttachAvailabilityChangeEvent"},
             {24, nullptr, "RecreateApplicationArea"},
         };
         RegisterHandlers(functions);
 
         auto& kernel = Core::System::GetInstance().Kernel();
-        activate_event =
-            Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "IUser:ActivateEvent");
         deactivate_event =
             Kernel::Event::Create(kernel, Kernel::ResetType::OneShot, "IUser:DeactivateEvent");
         availability_change_event = Kernel::Event::Create(kernel, Kernel::ResetType::OneShot,
@@ -59,6 +70,17 @@ public:
     }
 
 private:
+    struct TagInfo {
+        std::array<u8, 10> uuid;
+        u8 uuid_length; // TODO(ogniK): Figure out if this is actual the uuid length or does it
+                        // mean something else
+        INSERT_PADDING_BYTES(0x15);
+        u32_le protocol;
+        u32_le tag_type;
+        INSERT_PADDING_BYTES(0x2c);
+    };
+    static_assert(sizeof(TagInfo) == 0x54, "TagInfo is an invalid size");
+
     enum class State : u32 {
         NonInitialized = 0,
         Initialized = 1,
@@ -66,15 +88,40 @@ private:
 
     enum class DeviceState : u32 {
         Initialized = 0,
+        SearchingForTag = 1,
+        TagFound = 2,
+        TagRemoved = 3,
+        TagNearby = 4,
+        Unknown5 = 5,
+        Finalized = 6
     };
 
+    struct CommonInfo {
+        u16_be last_write_year;
+        u8 last_write_month;
+        u8 last_write_day;
+        u16_be write_counter;
+        u16_be version;
+        u32_be application_area_size;
+        INSERT_PADDING_BYTES(0x34);
+    };
+    static_assert(sizeof(CommonInfo) == 0x40, "CommonInfo is an invalid size");
+
     void Initialize(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_NFP, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2, 0};
+        rb.Push(RESULT_SUCCESS);
 
         state = State::Initialized;
 
-        IPC::ResponseBuilder rb{ctx, 2};
+        LOG_DEBUG(Service_NFC, "called");
+    }
+
+    void GetState(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 3, 0};
         rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u32>(static_cast<u32>(state));
+
+        LOG_DEBUG(Service_NFC, "called");
     }
 
     void ListDevices(Kernel::HLERequestContext& ctx) {
@@ -83,80 +130,219 @@ private:
 
         ctx.WriteBuffer(&device_handle, sizeof(device_handle));
 
-        LOG_WARNING(Service_NFP, "(STUBBED) called, array_size={}", array_size);
+        LOG_DEBUG(Service_NFP, "called, array_size={}", array_size);
 
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(0);
+        rb.Push<u32>(1);
+    }
+
+    void GetNpadId(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const u64 dev_handle = rp.Pop<u64>();
+        LOG_DEBUG(Service_NFP, "called, dev_handle=0x{:X}", dev_handle);
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(npad_id);
     }
 
     void AttachActivateEvent(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
         const u64 dev_handle = rp.Pop<u64>();
-        LOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
-
+        LOG_DEBUG(Service_NFP, "called, dev_handle=0x{:X}", dev_handle);
         IPC::ResponseBuilder rb{ctx, 2, 1};
         rb.Push(RESULT_SUCCESS);
-        rb.PushCopyObjects(activate_event);
+        rb.PushCopyObjects(nfp_interface.GetNFCEvent());
+        has_attached_handle = true;
     }
 
     void AttachDeactivateEvent(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
         const u64 dev_handle = rp.Pop<u64>();
-        LOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
+        LOG_DEBUG(Service_NFP, "called, dev_handle=0x{:X}", dev_handle);
 
         IPC::ResponseBuilder rb{ctx, 2, 1};
         rb.Push(RESULT_SUCCESS);
         rb.PushCopyObjects(deactivate_event);
     }
 
-    void GetState(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_NFP, "(STUBBED) called");
-        IPC::ResponseBuilder rb{ctx, 3};
+    void StopDetection(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+        switch (device_state) {
+        case DeviceState::TagFound:
+        case DeviceState::TagNearby:
+            deactivate_event->Signal();
+            device_state = DeviceState::Initialized;
+            break;
+        case DeviceState::SearchingForTag:
+        case DeviceState::TagRemoved:
+            device_state = DeviceState::Initialized;
+            break;
+        }
+        IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(static_cast<u32>(state));
     }
 
     void GetDeviceState(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_NFP, "(STUBBED) called");
+        LOG_DEBUG(Service_NFP, "called");
+        auto nfc_event = nfp_interface.GetNFCEvent();
+        if (!nfc_event->ShouldWait(Kernel::GetCurrentThread()) && !has_attached_handle) {
+            device_state = DeviceState::TagFound;
+            nfc_event->Clear();
+        }
+
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
         rb.Push<u32>(static_cast<u32>(device_state));
     }
 
-    void GetNpadId(Kernel::HLERequestContext& ctx) {
-        IPC::RequestParser rp{ctx};
-        const u64 dev_handle = rp.Pop<u64>();
-        LOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
-        IPC::ResponseBuilder rb{ctx, 3};
+    void StartDetection(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        if (device_state == DeviceState::Initialized || device_state == DeviceState::TagRemoved) {
+            device_state = DeviceState::SearchingForTag;
+        }
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetTagInfo(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        auto amiibo = nfp_interface.GetAmiiboBuffer();
+        TagInfo tag_info{};
+        std::memcpy(tag_info.uuid.data(), amiibo.uuid.data(), sizeof(tag_info.uuid.size()));
+        tag_info.uuid_length = static_cast<u8>(tag_info.uuid.size());
+
+        tag_info.protocol = 1; // TODO(ogniK): Figure out actual values
+        tag_info.tag_type = 2;
+        ctx.WriteBuffer(&tag_info, sizeof(TagInfo));
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void Mount(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        device_state = DeviceState::TagNearby;
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetModelInfo(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        auto amiibo = nfp_interface.GetAmiiboBuffer();
+        ctx.WriteBuffer(&amiibo.model_info, sizeof(amiibo.model_info));
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void Unmount(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        device_state = DeviceState::TagFound;
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void Finalize(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        device_state = DeviceState::Finalized;
+
+        IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(npad_id);
     }
 
     void AttachAvailabilityChangeEvent(Kernel::HLERequestContext& ctx) {
-        IPC::RequestParser rp{ctx};
-        const u64 dev_handle = rp.Pop<u64>();
-        LOG_WARNING(Service_NFP, "(STUBBED) called, dev_handle=0x{:X}", dev_handle);
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
 
         IPC::ResponseBuilder rb{ctx, 2, 1};
         rb.Push(RESULT_SUCCESS);
         rb.PushCopyObjects(availability_change_event);
     }
 
-    const u64 device_handle{0xDEAD};
-    const u32 npad_id{0}; // This is the first player controller id
+    void GetRegisterInfo(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
+
+        // TODO(ogniK): Pull Mii and owner data from amiibo
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetCommonInfo(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
+
+        // TODO(ogniK): Pull common information from amiibo
+
+        CommonInfo common_info{};
+        common_info.application_area_size = 0;
+        ctx.WriteBuffer(&common_info, sizeof(CommonInfo));
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void OpenApplicationArea(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+        // We don't need to worry about this since we can just open the file
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetApplicationAreaSize(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
+        // We don't need to worry about this since we can just open the file
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u32>(0); // This is from the GetCommonInfo stub
+    }
+
+    void GetApplicationArea(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NFP, "(STUBBED) called");
+
+        // TODO(ogniK): Pull application area from amiibo
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u32>(0); // This is from the GetCommonInfo stub
+    }
+
+    bool has_attached_handle{};
+    const u64 device_handle{Common::MakeMagic('Y', 'U', 'Z', 'U')};
+    const u32 npad_id{0}; // Player 1 controller
     State state{State::NonInitialized};
     DeviceState device_state{DeviceState::Initialized};
-    Kernel::SharedPtr<Kernel::Event> activate_event;
     Kernel::SharedPtr<Kernel::Event> deactivate_event;
     Kernel::SharedPtr<Kernel::Event> availability_change_event;
+    const Module::Interface& nfp_interface;
 };
 
 void Module::Interface::CreateUserInterface(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_NFP, "called");
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};
     rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<IUser>();
+    rb.PushIpcInterface<IUser>(*this);
+}
+
+bool Module::Interface::LoadAmiibo(const std::vector<u8>& buffer) {
+    std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
+    if (buffer.size() < sizeof(AmiiboFile)) {
+        return false;
+    }
+
+    std::memcpy(&amiibo, buffer.data(), sizeof(amiibo));
+    nfc_tag_load->Signal();
+    return true;
+}
+const Kernel::SharedPtr<Kernel::Event>& Module::Interface::GetNFCEvent() const {
+    return nfc_tag_load;
+}
+const Module::Interface::AmiiboFile& Module::Interface::GetAmiiboBuffer() const {
+    return amiibo;
 }
 
 void InstallInterfaces(SM::ServiceManager& service_manager) {

src/core/hle/service/nfp/nfp.h

@@ -4,6 +4,9 @@
 
 #pragma once
 
+#include <array>
+#include <vector>
+#include "core/hle/kernel/event.h"
 #include "core/hle/service/service.h"
 
 namespace Service::NFP {
@@ -15,7 +18,27 @@ public:
         explicit Interface(std::shared_ptr<Module> module, const char* name);
         ~Interface() override;
 
+        struct ModelInfo {
+            std::array<u8, 0x8> amiibo_identification_block;
+            INSERT_PADDING_BYTES(0x38);
+        };
+        static_assert(sizeof(ModelInfo) == 0x40, "ModelInfo is an invalid size");
+
+        struct AmiiboFile {
+            std::array<u8, 10> uuid;
+            INSERT_PADDING_BYTES(0x4a);
+            ModelInfo model_info;
+        };
+        static_assert(sizeof(AmiiboFile) == 0x94, "AmiiboFile is an invalid size");
+
         void CreateUserInterface(Kernel::HLERequestContext& ctx);
+        bool LoadAmiibo(const std::vector<u8>& buffer);
+        const Kernel::SharedPtr<Kernel::Event>& GetNFCEvent() const;
+        const AmiiboFile& GetAmiiboBuffer() const;
+
+    private:
+        Kernel::SharedPtr<Kernel::Event> nfc_tag_load{};
+        AmiiboFile amiibo{};
 
     protected:
         std::shared_ptr<Module> module;

src/core/hle/service/npns/npns.cpp

@@ -0,0 +1,88 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <memory>
+
+#include "core/hle/service/npns/npns.h"
+#include "core/hle/service/service.h"
+#include "core/hle/service/sm/sm.h"
+
+namespace Service::NPNS {
+
+class NPNS_S final : public ServiceFramework<NPNS_S> {
+public:
+    explicit NPNS_S() : ServiceFramework{"npns:s"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {1, nullptr, "ListenAll"},
+            {2, nullptr, "ListenTo"},
+            {3, nullptr, "Receive"},
+            {4, nullptr, "ReceiveRaw"},
+            {5, nullptr, "GetReceiveEvent"},
+            {6, nullptr, "ListenUndelivered"},
+            {7, nullptr, "GetStateChangeEVent"},
+            {11, nullptr, "SubscribeTopic"},
+            {12, nullptr, "UnsubscribeTopic"},
+            {13, nullptr, "QueryIsTopicExist"},
+            {21, nullptr, "CreateToken"},
+            {22, nullptr, "CreateTokenWithApplicationId"},
+            {23, nullptr, "DestroyToken"},
+            {24, nullptr, "DestroyTokenWithApplicationId"},
+            {25, nullptr, "QueryIsTokenValid"},
+            {31, nullptr, "UploadTokenToBaaS"},
+            {32, nullptr, "DestroyTokenForBaaS"},
+            {33, nullptr, "CreateTokenForBaaS"},
+            {34, nullptr, "SetBaaSDeviceAccountIdList"},
+            {101, nullptr, "Suspend"},
+            {102, nullptr, "Resume"},
+            {103, nullptr, "GetState"},
+            {104, nullptr, "GetStatistics"},
+            {105, nullptr, "GetPlayReportRequestEvent"},
+            {111, nullptr, "GetJid"},
+            {112, nullptr, "CreateJid"},
+            {113, nullptr, "DestroyJid"},
+            {114, nullptr, "AttachJid"},
+            {115, nullptr, "DetachJid"},
+            {201, nullptr, "RequestChangeStateForceTimed"},
+            {102, nullptr, "RequestChangeStateForceAsync"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
+class NPNS_U final : public ServiceFramework<NPNS_U> {
+public:
+    explicit NPNS_U() : ServiceFramework{"npns:u"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {1, nullptr, "ListenAll"},
+            {2, nullptr, "ListenTo"},
+            {3, nullptr, "Receive"},
+            {4, nullptr, "ReceiveRaw"},
+            {5, nullptr, "GetReceiveEvent"},
+            {7, nullptr, "GetStateChangeEVent"},
+            {21, nullptr, "CreateToken"},
+            {23, nullptr, "DestroyToken"},
+            {25, nullptr, "QueryIsTokenValid"},
+            {26, nullptr, "ListenToMyApplicationId"},
+            {101, nullptr, "Suspend"},
+            {102, nullptr, "Resume"},
+            {103, nullptr, "GetState"},
+            {104, nullptr, "GetStatistics"},
+            {111, nullptr, "GetJid"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
+void InstallInterfaces(SM::ServiceManager& sm) {
+    std::make_shared<NPNS_S>()->InstallAsService(sm);
+    std::make_shared<NPNS_U>()->InstallAsService(sm);
+}
+
+} // namespace Service::NPNS

src/core/hle/service/npns/npns.h

@@ -0,0 +1,15 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+namespace Service::SM {
+class ServiceManager;
+}
+
+namespace Service::NPNS {
+
+void InstallInterfaces(SM::ServiceManager& sm);
+
+} // namespace Service::NPNS

src/core/hle/service/prepo/prepo.cpp

@@ -14,20 +14,24 @@ public:
     explicit PlayReport(const char* name) : ServiceFramework{name} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {10100, nullptr, "SaveReport"},
-            {10101, &PlayReport::SaveReportWithUser, "SaveReportWithUser"},
+            {10100, nullptr, "SaveReportOld"},
+            {10101, &PlayReport::SaveReportWithUserOld, "SaveReportWithUserOld"},
+            {10102, nullptr, "SaveReport"},
+            {10103, nullptr, "SaveReportWithUser"},
             {10200, nullptr, "RequestImmediateTransmission"},
             {10300, nullptr, "GetTransmissionStatus"},
             {20100, nullptr, "SaveSystemReport"},
-            {20200, nullptr, "SetOperationMode"},
             {20101, nullptr, "SaveSystemReportWithUser"},
+            {20200, nullptr, "SetOperationMode"},
             {30100, nullptr, "ClearStorage"},
+            {30200, nullptr, "ClearStatistics"},
+            {30300, nullptr, "GetStorageUsage"},
+            {30400, nullptr, "GetStatistics"},
+            {30401, nullptr, "GetThroughputHistory"},
+            {30500, nullptr, "GetLastUploadError"},
             {40100, nullptr, "IsUserAgreementCheckEnabled"},
             {40101, nullptr, "SetUserAgreementCheckEnabled"},
-            {90100, nullptr, "GetStorageUsage"},
-            {90200, nullptr, "GetStatistics"},
-            {90201, nullptr, "GetThroughputHistory"},
-            {90300, nullptr, "GetLastUploadError"},
+            {90100, nullptr, "ReadAllReportFiles"},
         };
         // clang-format on
 
@@ -35,7 +39,7 @@ public:
     }
 
 private:
-    void SaveReportWithUser(Kernel::HLERequestContext& ctx) {
+    void SaveReportWithUserOld(Kernel::HLERequestContext& ctx) {
         // TODO(ogniK): Do we want to add play report?
         LOG_WARNING(Service_PREPO, "(STUBBED) called");
 
@@ -46,6 +50,7 @@ private:
 
 void InstallInterfaces(SM::ServiceManager& service_manager) {
     std::make_shared<PlayReport>("prepo:a")->InstallAsService(service_manager);
+    std::make_shared<PlayReport>("prepo:a2")->InstallAsService(service_manager);
     std::make_shared<PlayReport>("prepo:m")->InstallAsService(service_manager);
     std::make_shared<PlayReport>("prepo:s")->InstallAsService(service_manager);
     std::make_shared<PlayReport>("prepo:u")->InstallAsService(service_manager);

src/core/hle/service/service.cpp

@@ -22,7 +22,7 @@
 #include "core/hle/service/apm/apm.h"
 #include "core/hle/service/arp/arp.h"
 #include "core/hle/service/audio/audio.h"
-#include "core/hle/service/bcat/bcat.h"
+#include "core/hle/service/bcat/module.h"
 #include "core/hle/service/bpc/bpc.h"
 #include "core/hle/service/btdrv/btdrv.h"
 #include "core/hle/service/btm/btm.h"
@@ -48,11 +48,12 @@
 #include "core/hle/service/nfp/nfp.h"
 #include "core/hle/service/nifm/nifm.h"
 #include "core/hle/service/nim/nim.h"
+#include "core/hle/service/npns/npns.h"
 #include "core/hle/service/ns/ns.h"
 #include "core/hle/service/nvdrv/nvdrv.h"
 #include "core/hle/service/nvflinger/nvflinger.h"
 #include "core/hle/service/pcie/pcie.h"
-#include "core/hle/service/pctl/pctl.h"
+#include "core/hle/service/pctl/module.h"
 #include "core/hle/service/pcv/pcv.h"
 #include "core/hle/service/pm/pm.h"
 #include "core/hle/service/prepo/prepo.h"
@@ -237,6 +238,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, FileSys::VfsFilesystem& vfs)
     NFP::InstallInterfaces(*sm);
     NIFM::InstallInterfaces(*sm);
     NIM::InstallInterfaces(*sm);
+    NPNS::InstallInterfaces(*sm);
     NS::InstallInterfaces(*sm);
     Nvidia::InstallInterfaces(*sm, *nv_flinger);
     PCIe::InstallInterfaces(*sm);

src/core/hle/service/usb/usb.cpp

@@ -67,15 +67,15 @@ public:
     explicit IClientEpSession() : ServiceFramework{"IClientEpSession"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "Unknown1"},
-            {1, nullptr, "Unknown2"},
-            {2, nullptr, "Unknown3"},
-            {3, nullptr, "Unknown4"},
+            {0, nullptr, "Open"},
+            {1, nullptr, "Close"},
+            {2, nullptr, "Unknown1"},
+            {3, nullptr, "Populate"},
             {4, nullptr, "PostBufferAsync"},
-            {5, nullptr, "Unknown5"},
-            {6, nullptr, "Unknown6"},
-            {7, nullptr, "Unknown7"},
-            {8, nullptr, "Unknown8"},
+            {5, nullptr, "GetXferReport"},
+            {6, nullptr, "Unknown2"},
+            {7, nullptr, "Unknown3"},
+            {8, nullptr, "Unknown4"},
         };
         // clang-format on
 
@@ -89,15 +89,15 @@ public:
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "Unknown1"},
-            {1, nullptr, "Unknown2"},
-            {2, nullptr, "Unknown3"},
-            {3, nullptr, "Unknown4"},
-            {4, nullptr, "Unknown5"},
+            {1, nullptr, "SetInterface"},
+            {2, nullptr, "GetInterface"},
+            {3, nullptr, "GetAlternateInterface"},
+            {4, nullptr, "GetCurrentFrame"},
             {5, nullptr, "CtrlXferAsync"},
-            {6, nullptr, "Unknown6"},
+            {6, nullptr, "Unknown2"},
             {7, nullptr, "GetCtrlXferReport"},
-            {8, nullptr, "Unknown7"},
-            {9, nullptr, "GetClientEpSession"},
+            {8, nullptr, "ResetDevice"},
+            {9, nullptr, "OpenUsbEp"},
         };
         // clang-format on
 
@@ -111,13 +111,14 @@ public:
         // clang-format off
         static const FunctionInfo functions[] = {
             {0, nullptr, "BindClientProcess"},
-            {1, nullptr, "Unknown1"},
-            {2, nullptr, "Unknown2"},
-            {3, nullptr, "Unknown3"},
-            {4, nullptr, "Unknown4"},
-            {5, nullptr, "Unknown5"},
+            {1, nullptr, "QueryAllInterfaces"},
+            {2, nullptr, "QueryAvailableInterfaces"},
+            {3, nullptr, "QueryAcquiredInterfaces"},
+            {4, nullptr, "CreateInterfaceAvailableEvent"},
+            {5, nullptr, "DestroyInterfaceAvailableEvent"},
             {6, nullptr, "GetInterfaceStateChangeEvent"},
-            {7, nullptr, "GetClientIfSession"},
+            {7, nullptr, "AcquireUsbIf"},
+            {8, nullptr, "Unknown1"},
         };
         // clang-format on
 

src/core/loader/nro.cpp

@@ -127,18 +127,23 @@ static constexpr u32 PageAlignSize(u32 size) {
     return (size + Memory::PAGE_MASK) & ~Memory::PAGE_MASK;
 }
 
-bool AppLoader_NRO::LoadNro(const FileSys::VfsFile& file, VAddr load_base) {
-    // Read NSO header
-    NroHeader nro_header{};
-    if (sizeof(NroHeader) != file.ReadObject(&nro_header)) {
+/*static*/ bool AppLoader_NRO::LoadNro(const std::vector<u8>& data, const std::string& name,
+                                       VAddr load_base) {
+
+    if (data.size() < sizeof(NroHeader)) {
         return {};
     }
+
+    // Read NSO header
+    NroHeader nro_header{};
+    std::memcpy(&nro_header, data.data(), sizeof(NroHeader));
     if (nro_header.magic != Common::MakeMagic('N', 'R', 'O', '0')) {
         return {};
     }
 
     // Build program image
-    std::vector<u8> program_image = file.ReadBytes(PageAlignSize(nro_header.file_size));
+    std::vector<u8> program_image(PageAlignSize(nro_header.file_size));
+    std::memcpy(program_image.data(), data.data(), program_image.size());
     if (program_image.size() != PageAlignSize(nro_header.file_size)) {
         return {};
     }
@@ -182,11 +187,15 @@ bool AppLoader_NRO::LoadNro(const FileSys::VfsFile& file, VAddr load_base) {
     Core::CurrentProcess()->LoadModule(std::move(codeset), load_base);
 
     // Register module with GDBStub
-    GDBStub::RegisterModule(file.GetName(), load_base, load_base);
+    GDBStub::RegisterModule(name, load_base, load_base);
 
     return true;
 }
 
+bool AppLoader_NRO::LoadNro(const FileSys::VfsFile& file, VAddr load_base) {
+    return AppLoader_NRO::LoadNro(file.ReadAllBytes(), file.GetName(), load_base);
+}
+
 ResultStatus AppLoader_NRO::Load(Kernel::Process& process) {
     if (is_loaded) {
         return ResultStatus::ErrorAlreadyLoaded;

src/core/loader/nro.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <string>
+#include <vector>
 #include "common/common_types.h"
 #include "core/loader/linker.h"
 #include "core/loader/loader.h"
@@ -40,6 +41,8 @@ public:
     ResultStatus ReadTitle(std::string& title) override;
     bool IsRomFSUpdatable() const override;
 
+    static bool LoadNro(const std::vector<u8>& data, const std::string& name, VAddr load_base);
+
 private:
     bool LoadNro(const FileSys::VfsFile& file, VAddr load_base);
 

src/core/loader/nsp.cpp

@@ -36,6 +36,16 @@ AppLoader_NSP::AppLoader_NSP(FileSys::VirtualFile file)
 
     std::tie(nacp_file, icon_file) =
         FileSys::PatchManager(nsp->GetProgramTitleID()).ParseControlNCA(*control_nca);
+
+    if (nsp->IsExtractedType()) {
+        secondary_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(nsp->GetExeFS());
+    } else {
+        if (title_id == 0)
+            return;
+
+        secondary_loader = std::make_unique<AppLoader_NCA>(
+            nsp->GetNCAFile(title_id, FileSys::ContentRecordType::Program));
+    }
 }
 
 AppLoader_NSP::~AppLoader_NSP() = default;
@@ -67,26 +77,19 @@ ResultStatus AppLoader_NSP::Load(Kernel::Process& process) {
         return ResultStatus::ErrorAlreadyLoaded;
     }
 
-    if (nsp->IsExtractedType()) {
-        secondary_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(nsp->GetExeFS());
-    } else {
-        if (title_id == 0)
-            return ResultStatus::ErrorNSPMissingProgramNCA;
+    if (title_id == 0)
+        return ResultStatus::ErrorNSPMissingProgramNCA;
 
-        secondary_loader = std::make_unique<AppLoader_NCA>(
-            nsp->GetNCAFile(title_id, FileSys::ContentRecordType::Program));
+    if (nsp->GetStatus() != ResultStatus::Success)
+        return nsp->GetStatus();
 
-        if (nsp->GetStatus() != ResultStatus::Success)
-            return nsp->GetStatus();
+    if (nsp->GetProgramStatus(title_id) != ResultStatus::Success)
+        return nsp->GetProgramStatus(title_id);
 
-        if (nsp->GetProgramStatus(title_id) != ResultStatus::Success)
-            return nsp->GetProgramStatus(title_id);
-
-        if (nsp->GetNCA(title_id, FileSys::ContentRecordType::Program) == nullptr) {
-            if (!Core::Crypto::KeyManager::KeyFileExists(false))
-                return ResultStatus::ErrorMissingProductionKeyFile;
-            return ResultStatus::ErrorNSPMissingProgramNCA;
-        }
+    if (nsp->GetNCA(title_id, FileSys::ContentRecordType::Program) == nullptr) {
+        if (!Core::Crypto::KeyManager::KeyFileExists(false))
+            return ResultStatus::ErrorMissingProductionKeyFile;
+        return ResultStatus::ErrorNSPMissingProgramNCA;
     }
 
     const auto result = secondary_loader->Load(process);

src/core/perf_stats.cpp

@@ -74,10 +74,6 @@ double PerfStats::GetLastFrameTimeScale() {
 }
 
 void FrameLimiter::DoFrameLimiting(microseconds current_system_time_us) {
-    // Max lag caused by slow frames. Can be adjusted to compensate for too many slow frames. Higher
-    // values increase the time needed to recover and limit framerate again after spikes.
-    constexpr microseconds MAX_LAG_TIME_US = 25000us;
-
     if (!Settings::values.use_frame_limit) {
         return;
     }

src/core/settings.h

@@ -113,7 +113,8 @@ static const std::array<const char*, NumAnalogs> mapping = {{
 struct Values {
     // System
     bool use_docked_mode;
-    std::string username;
+    bool enable_nfc;
+    int current_user;
     int language_index;
 
     // Controls

src/video_core/engines/maxwell_3d.cpp

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <cinttypes>
+#include <cstring>
 #include "common/assert.h"
 #include "core/core.h"
 #include "core/core_timing.h"
@@ -13,14 +14,30 @@
 #include "video_core/renderer_base.h"
 #include "video_core/textures/texture.h"
 
-namespace Tegra {
-namespace Engines {
+namespace Tegra::Engines {
 
 /// First register id that is actually a Macro call.
 constexpr u32 MacroRegistersStart = 0xE00;
 
 Maxwell3D::Maxwell3D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager)
-    : memory_manager(memory_manager), rasterizer{rasterizer}, macro_interpreter(*this) {}
+    : memory_manager(memory_manager), rasterizer{rasterizer}, macro_interpreter(*this) {
+    InitializeRegisterDefaults();
+}
+
+void Maxwell3D::InitializeRegisterDefaults() {
+    // Initializes registers to their default values - what games expect them to be at boot. This is
+    // for certain registers that may not be explicitly set by games.
+
+    // Reset all registers to zero
+    std::memset(&regs, 0, sizeof(regs));
+
+    // Depth range near/far is not always set, but is expected to be the default 0.0f, 1.0f. This is
+    // needed for ARMS.
+    for (std::size_t viewport{}; viewport < Regs::NumViewports; ++viewport) {
+        regs.viewport[viewport].depth_range_near = 0.0f;
+        regs.viewport[viewport].depth_range_far = 1.0f;
+    }
+}
 
 void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
     // Reset the current macro.
@@ -156,7 +173,6 @@ void Maxwell3D::ProcessQueryGet() {
     ASSERT_MSG(regs.query.query_get.unit == Regs::QueryUnit::Crop,
                "Units other than CROP are unimplemented");
 
-    u32 value = Memory::Read32(*address);
     u64 result = 0;
 
     // TODO(Subv): Support the other query variables
@@ -408,5 +424,4 @@ void Maxwell3D::ProcessClearBuffers() {
     rasterizer.Clear();
 }
 
-} // namespace Engines
-} // namespace Tegra
+} // namespace Tegra::Engines

src/video_core/engines/maxwell_3d.h

@@ -643,8 +643,10 @@ public:
                 u32 d3d_cull_mode;
 
                 ComparisonOp depth_test_func;
+                float alpha_test_ref;
+                ComparisonOp alpha_test_func;
 
-                INSERT_PADDING_WORDS(0xB);
+                INSERT_PADDING_WORDS(0x9);
 
                 struct {
                     u32 separate_alpha;
@@ -982,6 +984,8 @@ public:
     Texture::FullTextureInfo GetStageTexture(Regs::ShaderStage stage, std::size_t offset) const;
 
 private:
+    void InitializeRegisterDefaults();
+
     VideoCore::RasterizerInterface& rasterizer;
 
     std::unordered_map<u32, std::vector<u32>> uploaded_macros;

src/video_core/engines/maxwell_compute.cpp

@@ -6,8 +6,7 @@
 #include "core/core.h"
 #include "video_core/engines/maxwell_compute.h"
 
-namespace Tegra {
-namespace Engines {
+namespace Tegra::Engines {
 
 void MaxwellCompute::WriteReg(u32 method, u32 value) {
     ASSERT_MSG(method < Regs::NUM_REGS,
@@ -26,5 +25,4 @@ void MaxwellCompute::WriteReg(u32 method, u32 value) {
     }
 }
 
-} // namespace Engines
-} // namespace Tegra
+} // namespace Tegra::Engines

src/video_core/engines/maxwell_dma.cpp

@@ -7,8 +7,7 @@
 #include "video_core/rasterizer_interface.h"
 #include "video_core/textures/decoders.h"
 
-namespace Tegra {
-namespace Engines {
+namespace Tegra::Engines {
 
 MaxwellDMA::MaxwellDMA(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager)
     : memory_manager(memory_manager), rasterizer{rasterizer} {}
@@ -78,9 +77,9 @@ void MaxwellDMA::HandleCopy() {
 
     ASSERT(regs.exec.enable_2d == 1);
 
-    std::size_t copy_size = regs.x_count * regs.y_count;
+    const std::size_t copy_size = regs.x_count * regs.y_count;
 
-    const auto FlushAndInvalidate = [&](u32 src_size, u32 dst_size) {
+    const auto FlushAndInvalidate = [&](u32 src_size, u64 dst_size) {
         // TODO(Subv): For now, manually flush the regions until we implement GPU-accelerated
         // copying.
         rasterizer.FlushRegion(source_cpu, src_size);
@@ -91,14 +90,11 @@ void MaxwellDMA::HandleCopy() {
         rasterizer.InvalidateRegion(dest_cpu, dst_size);
     };
 
-    u8* src_buffer = Memory::GetPointer(source_cpu);
-    u8* dst_buffer = Memory::GetPointer(dest_cpu);
-
     if (regs.exec.is_dst_linear && !regs.exec.is_src_linear) {
         ASSERT(regs.src_params.size_z == 1);
         // If the input is tiled and the output is linear, deswizzle the input and copy it over.
 
-        u32 src_bytes_per_pixel = regs.src_pitch / regs.src_params.size_x;
+        const u32 src_bytes_per_pixel = regs.src_pitch / regs.src_params.size_x;
 
         FlushAndInvalidate(regs.src_pitch * regs.src_params.size_y,
                            copy_size * src_bytes_per_pixel);
@@ -111,7 +107,7 @@ void MaxwellDMA::HandleCopy() {
         ASSERT(regs.dst_params.size_z == 1);
         ASSERT(regs.src_pitch == regs.x_count);
 
-        u32 src_bpp = regs.src_pitch / regs.x_count;
+        const u32 src_bpp = regs.src_pitch / regs.x_count;
 
         FlushAndInvalidate(regs.src_pitch * regs.y_count,
                            regs.dst_params.size_x * regs.dst_params.size_y * src_bpp);
@@ -122,5 +118,4 @@ void MaxwellDMA::HandleCopy() {
     }
 }
 
-} // namespace Engines
-} // namespace Tegra
+} // namespace Tegra::Engines

src/video_core/engines/shader_bytecode.h

@@ -79,6 +79,7 @@ union Attribute {
     constexpr explicit Attribute(u64 value) : value(value) {}
 
     enum class Index : u64 {
+        PointSize = 6,
         Position = 7,
         Attribute_0 = 8,
         Attribute_31 = 39,
@@ -207,6 +208,16 @@ enum class UniformType : u64 {
     Double = 5,
 };
 
+enum class StoreType : u64 {
+    Unsigned8 = 0,
+    Signed8 = 1,
+    Unsigned16 = 2,
+    Signed16 = 3,
+    Bytes32 = 4,
+    Bytes64 = 5,
+    Bytes128 = 6,
+};
+
 enum class IMinMaxExchange : u64 {
     None = 0,
     XLo = 1,
@@ -214,7 +225,7 @@ enum class IMinMaxExchange : u64 {
     XHi = 3,
 };
 
-enum class VmadType : u64 {
+enum class VideoType : u64 {
     Size16_Low = 0,
     Size16_High = 1,
     Size32 = 2,
@@ -563,6 +574,10 @@ union Instruction {
         BitField<48, 1, u64> negate_b;
     } fmul;
 
+    union {
+        BitField<55, 1, u64> saturate;
+    } fmul32;
+
     union {
         BitField<48, 1, u64> is_signed;
     } shift;
@@ -742,6 +757,18 @@ union Instruction {
         BitField<44, 2, u64> unknown;
     } ld_c;
 
+    union {
+        BitField<48, 3, StoreType> type;
+    } ldst_sl;
+
+    union {
+        BitField<44, 2, u64> unknown;
+    } ld_l;
+
+    union {
+        BitField<44, 2, u64> unknown;
+    } st_l;
+
     union {
         BitField<0, 3, u64> pred0;
         BitField<3, 3, u64> pred3;
@@ -753,7 +780,6 @@ union Instruction {
         BitField<45, 2, PredOperation> op;
         BitField<47, 1, u64> ftz;
         BitField<48, 4, PredCondition> cond;
-        BitField<56, 1, u64> neg_b;
     } fsetp;
 
     union {
@@ -779,6 +805,14 @@ union Instruction {
         BitField<45, 2, PredOperation> op;
     } psetp;
 
+    union {
+        BitField<43, 4, PredCondition> cond;
+        BitField<45, 2, PredOperation> op;
+        BitField<3, 3, u64> pred3;
+        BitField<0, 3, u64> pred0;
+        BitField<39, 3, u64> pred39;
+    } vsetp;
+
     union {
         BitField<12, 3, u64> pred12;
         BitField<15, 1, u64> neg_pred12;
@@ -828,7 +862,6 @@ union Instruction {
         BitField<53, 1, u64> neg_b;
         BitField<54, 1, u64> abs_a;
         BitField<55, 1, u64> ftz;
-        BitField<56, 1, u64> neg_imm;
     } fset;
 
     union {
@@ -1152,15 +1185,17 @@ union Instruction {
     union {
         BitField<48, 1, u64> signed_a;
         BitField<38, 1, u64> is_byte_chunk_a;
-        BitField<36, 2, VmadType> type_a;
+        BitField<36, 2, VideoType> type_a;
         BitField<36, 2, u64> byte_height_a;
 
         BitField<49, 1, u64> signed_b;
         BitField<50, 1, u64> use_register_b;
         BitField<30, 1, u64> is_byte_chunk_b;
-        BitField<28, 2, VmadType> type_b;
+        BitField<28, 2, VideoType> type_b;
         BitField<28, 2, u64> byte_height_b;
+    } video;
 
+    union {
         BitField<51, 2, VmadShr> shr;
         BitField<55, 1, u64> saturate; // Saturates the result (a * b + c)
         BitField<47, 1, u64> cc;
@@ -1196,6 +1231,7 @@ union Instruction {
     BitField<61, 1, u64> is_b_imm;
     BitField<60, 1, u64> is_b_gpr;
     BitField<59, 1, u64> is_c_gpr;
+    BitField<20, 24, s64> smem_imm;
 
     Attribute attribute;
     Sampler sampler;
@@ -1211,14 +1247,20 @@ public:
         KIL,
         SSY,
         SYNC,
+        BRK,
         DEPBAR,
         BFE_C,
         BFE_R,
         BFE_IMM,
         BRA,
+        PBK,
         LD_A,
+        LD_L,
+        LD_S,
         LD_C,
         ST_A,
+        ST_L,
+        ST_S,
         LDG, // Load from global memory
         STG, // Store in global memory
         TEX,
@@ -1234,6 +1276,7 @@ public:
         OUT_R, // Emit vertex/primitive
         ISBERD,
         VMAD,
+        VSETP,
         FFMA_IMM, // Fused Multiply and Add
         FFMA_CR,
         FFMA_RC,
@@ -1372,7 +1415,7 @@ public:
     /// conditionally executed).
     static bool IsPredicatedInstruction(Id opcode) {
         // TODO(Subv): Add the rest of unpredicated instructions.
-        return opcode != Id::SSY;
+        return opcode != Id::SSY && opcode != Id::PBK;
     }
 
     class Matcher {
@@ -1468,12 +1511,18 @@ private:
 #define INST(bitstring, op, type, name) Detail::GetMatcher(bitstring, op, type, name)
             INST("111000110011----", Id::KIL, Type::Flow, "KIL"),
             INST("111000101001----", Id::SSY, Type::Flow, "SSY"),
+            INST("111000101010----", Id::PBK, Type::Flow, "PBK"),
             INST("111000100100----", Id::BRA, Type::Flow, "BRA"),
+            INST("1111000011111---", Id::SYNC, Type::Flow, "SYNC"),
+            INST("111000110100---", Id::BRK, Type::Flow, "BRK"),
             INST("1111000011110---", Id::DEPBAR, Type::Synch, "DEPBAR"),
-            INST("1111000011111---", Id::SYNC, Type::Synch, "SYNC"),
             INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"),
+            INST("1110111101001---", Id::LD_S, Type::Memory, "LD_S"),
+            INST("1110111101000---", Id::LD_L, Type::Memory, "LD_L"),
             INST("1110111110010---", Id::LD_C, Type::Memory, "LD_C"),
             INST("1110111111110---", Id::ST_A, Type::Memory, "ST_A"),
+            INST("1110111101011---", Id::ST_S, Type::Memory, "ST_S"),
+            INST("1110111101010---", Id::ST_L, Type::Memory, "ST_L"),
             INST("1110111011010---", Id::LDG, Type::Memory, "LDG"),
             INST("1110111011011---", Id::STG, Type::Memory, "STG"),
             INST("110000----111---", Id::TEX, Type::Memory, "TEX"),
@@ -1489,6 +1538,7 @@ private:
             INST("1111101111100---", Id::OUT_R, Type::Trivial, "OUT_R"),
             INST("1110111111010---", Id::ISBERD, Type::Trivial, "ISBERD"),
             INST("01011111--------", Id::VMAD, Type::Trivial, "VMAD"),
+            INST("0101000011110---", Id::VSETP, Type::Trivial, "VSETP"),
             INST("0011001-1-------", Id::FFMA_IMM, Type::Ffma, "FFMA_IMM"),
             INST("010010011-------", Id::FFMA_CR, Type::Ffma, "FFMA_CR"),
             INST("010100011-------", Id::FFMA_RC, Type::Ffma, "FFMA_RC"),
@@ -1608,4 +1658,4 @@ private:
     }
 };
 
-} // namespace Tegra::Shader
+} // namespace Tegra::Shader

src/video_core/engines/shader_header.h

@@ -96,6 +96,11 @@ struct Header {
             }
         } ps;
     };
+
+    u64 GetLocalMemorySize() {
+        return (common1.shader_local_memory_low_size |
+                (common2.shader_local_memory_high_size << 24));
+    }
 };
 
 static_assert(sizeof(Header) == 0x50, "Incorrect structure size");

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -570,10 +570,12 @@ void RasterizerOpenGL::DrawArrays() {
     SyncBlendState();
     SyncLogicOpState();
     SyncCullMode();
-    SyncAlphaTest();
+    SyncDepthRange();
     SyncScissorTest();
+    // Alpha Testing is synced on shaders.
     SyncTransformFeedback();
     SyncPointState();
+    CheckAlphaTests();
 
     // TODO(bunnei): Sync framebuffer_scale uniform here
     // TODO(bunnei): Sync scissorbox uniform(s) here
@@ -922,12 +924,11 @@ void RasterizerOpenGL::SyncCullMode() {
     }
 }
 
-void RasterizerOpenGL::SyncDepthScale() {
-    UNREACHABLE();
-}
+void RasterizerOpenGL::SyncDepthRange() {
+    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
-void RasterizerOpenGL::SyncDepthOffset() {
-    UNREACHABLE();
+    state.depth.depth_range_near = regs.viewport->depth_range_near;
+    state.depth.depth_range_far = regs.viewport->depth_range_far;
 }
 
 void RasterizerOpenGL::SyncDepthTestState() {
@@ -1007,17 +1008,6 @@ void RasterizerOpenGL::SyncLogicOpState() {
     state.logic_op.operation = MaxwellToGL::LogicOp(regs.logic_op.operation);
 }
 
-void RasterizerOpenGL::SyncAlphaTest() {
-    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
-
-    // TODO(Rodrigo): Alpha testing is a legacy OpenGL feature, but it can be
-    // implemented with a test+discard in fragment shaders.
-    if (regs.alpha_test_enabled != 0) {
-        LOG_CRITICAL(Render_OpenGL, "Alpha testing is not implemented");
-        UNREACHABLE();
-    }
-}
-
 void RasterizerOpenGL::SyncScissorTest() {
     const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
 
@@ -1052,4 +1042,15 @@ void RasterizerOpenGL::SyncPointState() {
     state.point.size = regs.point_size == 0 ? 1 : regs.point_size;
 }
 
+void RasterizerOpenGL::CheckAlphaTests() {
+    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+
+    if (regs.alpha_test_enabled != 0 && regs.rt_control.count > 1) {
+        LOG_CRITICAL(
+            Render_OpenGL,
+            "Alpha Testing is enabled with Multiple Render Targets, this behavior is undefined.");
+        UNREACHABLE();
+    }
+}
+
 } // namespace OpenGL

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -144,11 +144,8 @@ private:
     /// Syncs the cull mode to match the guest state
     void SyncCullMode();
 
-    /// Syncs the depth scale to match the guest state
-    void SyncDepthScale();
-
-    /// Syncs the depth offset to match the guest state
-    void SyncDepthOffset();
+    /// Syncs the depth range to match the guest state
+    void SyncDepthRange();
 
     /// Syncs the depth test state to match the guest state
     void SyncDepthTestState();
@@ -162,9 +159,6 @@ private:
     /// Syncs the LogicOp state to match the guest state
     void SyncLogicOpState();
 
-    /// Syncs the alpha test state to match the guest state
-    void SyncAlphaTest();
-
     /// Syncs the scissor test state to match the guest state
     void SyncScissorTest();
 
@@ -174,6 +168,9 @@ private:
     /// Syncs the point state to match the guest state
     void SyncPointState();
 
+    /// Check asserts for alpha testing.
+    void CheckAlphaTests();
+
     bool has_ARB_direct_state_access = false;
     bool has_ARB_multi_bind = false;
     bool has_ARB_separate_shader_objects = false;

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -78,6 +78,29 @@ void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
     }
 }
 
+std::size_t SurfaceParams::InnerMemorySize(bool layer_only) const {
+    const u32 compression_factor{GetCompressionFactor(pixel_format)};
+    const u32 bytes_per_pixel{GetBytesPerPixel(pixel_format)};
+    u32 m_depth = (layer_only ? 1U : depth);
+    u32 m_width = std::max(1U, width / compression_factor);
+    u32 m_height = std::max(1U, height / compression_factor);
+    std::size_t size = Tegra::Texture::CalculateSize(is_tiled, bytes_per_pixel, m_width, m_height,
+                                                     m_depth, block_height, block_depth);
+    u32 m_block_height = block_height;
+    u32 m_block_depth = block_depth;
+    std::size_t block_size_bytes = 512 * block_height * block_depth; // 512 is GOB size
+    for (u32 i = 1; i < max_mip_level; i++) {
+        m_width = std::max(1U, m_width / 2);
+        m_height = std::max(1U, m_height / 2);
+        m_depth = std::max(1U, m_depth / 2);
+        m_block_height = std::max(1U, m_block_height / 2);
+        m_block_depth = std::max(1U, m_block_depth / 2);
+        size += Tegra::Texture::CalculateSize(is_tiled, bytes_per_pixel, m_width, m_height, m_depth,
+                                              m_block_height, m_block_depth);
+    }
+    return is_tiled ? Common::AlignUp(size, block_size_bytes) : size;
+}
+
 /*static*/ SurfaceParams SurfaceParams::CreateForTexture(
     const Tegra::Texture::FullTextureInfo& config, const GLShader::SamplerEntry& entry) {
     SurfaceParams params{};
@@ -124,6 +147,7 @@ void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
         break;
     }
 
+    params.is_layered = SurfaceTargetIsLayered(params.target);
     params.max_mip_level = config.tic.max_mip_level + 1;
     params.rt = {};
 
@@ -150,6 +174,7 @@ void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
     params.target = SurfaceTarget::Texture2D;
     params.depth = 1;
     params.max_mip_level = 0;
+    params.is_layered = false;
 
     // Render target specific parameters, not used for caching
     params.rt.index = static_cast<u32>(index);
@@ -182,6 +207,7 @@ void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
     params.target = SurfaceTarget::Texture2D;
     params.depth = 1;
     params.max_mip_level = 0;
+    params.is_layered = false;
     params.rt = {};
 
     params.InitCacheParameters(zeta_address);
@@ -361,10 +387,11 @@ void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 d
     }
 }
 
-static constexpr std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
-                            SurfaceParams::MaxPixelFormat>
-    morton_to_gl_fns = {
-        // clang-format off
+using GLConversionArray = std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
+                                     SurfaceParams::MaxPixelFormat>;
+
+static constexpr GLConversionArray morton_to_gl_fns = {
+    // clang-format off
         MortonCopy<true, PixelFormat::ABGR8U>,
         MortonCopy<true, PixelFormat::ABGR8S>,
         MortonCopy<true, PixelFormat::ABGR8UI>,
@@ -418,13 +445,11 @@ static constexpr std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t,
         MortonCopy<true, PixelFormat::Z24S8>,
         MortonCopy<true, PixelFormat::S8Z24>,
         MortonCopy<true, PixelFormat::Z32FS8>,
-        // clang-format on
+    // clang-format on
 };
 
-static constexpr std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
-                            SurfaceParams::MaxPixelFormat>
-    gl_to_morton_fns = {
-        // clang-format off
+static constexpr GLConversionArray gl_to_morton_fns = {
+    // clang-format off
         MortonCopy<false, PixelFormat::ABGR8U>,
         MortonCopy<false, PixelFormat::ABGR8S>,
         MortonCopy<false, PixelFormat::ABGR8UI>,
@@ -479,9 +504,35 @@ static constexpr std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t,
         MortonCopy<false, PixelFormat::Z24S8>,
         MortonCopy<false, PixelFormat::S8Z24>,
         MortonCopy<false, PixelFormat::Z32FS8>,
-        // clang-format on
+    // clang-format on
 };
 
+void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params,
+                 std::vector<u8>& gl_buffer) {
+    u32 depth = params.depth;
+    if (params.target == SurfaceParams::SurfaceTarget::Texture2D) {
+        // TODO(Blinkhawk): Eliminate this condition once all texture types are implemented.
+        depth = 1U;
+    }
+    if (params.is_layered) {
+        u64 offset = 0;
+        u64 offset_gl = 0;
+        u64 layer_size = params.LayerMemorySize();
+        u64 gl_size = params.LayerSizeGL();
+        for (u32 i = 0; i < depth; i++) {
+            functions[static_cast<std::size_t>(params.pixel_format)](
+                params.width, params.block_height, params.height, params.block_depth, 1,
+                gl_buffer.data() + offset_gl, gl_size, params.addr + offset);
+            offset += layer_size;
+            offset_gl += gl_size;
+        }
+    } else {
+        functions[static_cast<std::size_t>(params.pixel_format)](
+            params.width, params.block_height, params.height, params.block_depth, depth,
+            gl_buffer.data(), gl_buffer.size(), params.addr);
+    }
+}
+
 static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
                         GLuint read_fb_handle, GLuint draw_fb_handle, GLenum src_attachment = 0,
                         GLenum dst_attachment = 0, std::size_t cubemap_face = 0) {
@@ -881,21 +932,10 @@ void CachedSurface::LoadGLBuffer() {
 
     gl_buffer.resize(params.size_in_bytes_gl);
     if (params.is_tiled) {
-        u32 depth = params.depth;
-        u32 block_depth = params.block_depth;
-
         ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
                    params.block_width, static_cast<u32>(params.target));
 
-        if (params.target == SurfaceParams::SurfaceTarget::Texture2D) {
-            // TODO(Blinkhawk): Eliminate this condition once all texture types are implemented.
-            depth = 1U;
-            block_depth = 1U;
-        }
-
-        morton_to_gl_fns[static_cast<std::size_t>(params.pixel_format)](
-            params.width, params.block_height, params.height, block_depth, depth, gl_buffer.data(),
-            gl_buffer.size(), params.addr);
+        SwizzleFunc(morton_to_gl_fns, params, gl_buffer);
     } else {
         const auto texture_src_data{Memory::GetPointer(params.addr)};
         const auto texture_src_data_end{texture_src_data + params.size_in_bytes_gl};
@@ -929,19 +969,10 @@ void CachedSurface::FlushGLBuffer() {
     const u8* const texture_src_data = Memory::GetPointer(params.addr);
     ASSERT(texture_src_data);
     if (params.is_tiled) {
-        u32 depth = params.depth;
-        u32 block_depth = params.block_depth;
-
         ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
                    params.block_width, static_cast<u32>(params.target));
 
-        if (params.target == SurfaceParams::SurfaceTarget::Texture2D) {
-            // TODO(Blinkhawk): Eliminate this condition once all texture types are implemented.
-            depth = 1U;
-        }
-        gl_to_morton_fns[static_cast<size_t>(params.pixel_format)](
-            params.width, params.block_height, params.height, block_depth, depth, gl_buffer.data(),
-            gl_buffer.size(), GetAddr());
+        SwizzleFunc(gl_to_morton_fns, params, gl_buffer);
     } else {
         std::memcpy(Memory::GetPointer(GetAddr()), gl_buffer.data(), GetSizeInBytes());
     }
@@ -1179,7 +1210,7 @@ void RasterizerCacheOpenGL::AccurateCopySurface(const Surface& src_surface,
                                                 const Surface& dst_surface) {
     const auto& src_params{src_surface->GetSurfaceParams()};
     const auto& dst_params{dst_surface->GetSurfaceParams()};
-    FlushRegion(src_params.addr, dst_params.size_in_bytes);
+    FlushRegion(src_params.addr, dst_params.MemorySize());
     LoadSurface(dst_surface);
 }
 
@@ -1221,44 +1252,10 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
             CopySurface(old_surface, new_surface, copy_pbo.handle);
         }
         break;
+    case SurfaceParams::SurfaceTarget::TextureCubemap:
     case SurfaceParams::SurfaceTarget::Texture3D:
         AccurateCopySurface(old_surface, new_surface);
         break;
-    case SurfaceParams::SurfaceTarget::TextureCubemap: {
-        if (old_params.rt.array_mode != 1) {
-            // TODO(bunnei): This is used by Breath of the Wild, I'm not sure how to implement this
-            // yet (array rendering used as a cubemap texture).
-            LOG_CRITICAL(HW_GPU, "Unhandled rendertarget array_mode {}", old_params.rt.array_mode);
-            UNREACHABLE();
-            return new_surface;
-        }
-
-        // This seems to be used for render-to-cubemap texture
-        ASSERT_MSG(old_params.target == SurfaceParams::SurfaceTarget::Texture2D, "Unexpected");
-        ASSERT_MSG(old_params.pixel_format == new_params.pixel_format, "Unexpected");
-        ASSERT_MSG(old_params.rt.base_layer == 0, "Unimplemented");
-
-        // TODO(bunnei): Verify the below - this stride seems to be in 32-bit words, not pixels.
-        // Tested with Splatoon 2, Super Mario Odyssey, and Breath of the Wild.
-        const std::size_t byte_stride{old_params.rt.layer_stride * sizeof(u32)};
-
-        for (std::size_t index = 0; index < new_params.depth; ++index) {
-            Surface face_surface{TryGetReservedSurface(old_params)};
-            ASSERT_MSG(face_surface, "Unexpected");
-
-            if (is_blit) {
-                BlitSurface(face_surface, new_surface, read_framebuffer.handle,
-                            draw_framebuffer.handle, face_surface->GetSurfaceParams().rt.index,
-                            new_params.rt.index, index);
-            } else {
-                CopySurface(face_surface, new_surface, copy_pbo.handle,
-                            face_surface->GetSurfaceParams().rt.index, new_params.rt.index, index);
-            }
-
-            old_params.addr += byte_stride;
-        }
-        break;
-    }
     default:
         LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
                      static_cast<u32>(new_params.target));
@@ -1266,7 +1263,7 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
     }
 
     return new_surface;
-}
+} // namespace OpenGL
 
 Surface RasterizerCacheOpenGL::TryFindFramebufferSurface(VAddr addr) const {
     return TryGet(addr);

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -168,6 +168,23 @@ struct SurfaceParams {
         }
     }
 
+    static bool SurfaceTargetIsLayered(SurfaceTarget target) {
+        switch (target) {
+        case SurfaceTarget::Texture1D:
+        case SurfaceTarget::Texture2D:
+        case SurfaceTarget::Texture3D:
+            return false;
+        case SurfaceTarget::Texture1DArray:
+        case SurfaceTarget::Texture2DArray:
+        case SurfaceTarget::TextureCubemap:
+            return true;
+        default:
+            LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
+            UNREACHABLE();
+            return false;
+        }
+    }
+
     /**
      * Gets the compression factor for the specified PixelFormat. This applies to just the
      * "compressed width" and "compressed height", not the overall compression factor of a
@@ -742,6 +759,25 @@ struct SurfaceParams {
         return size_in_bytes_gl / 6;
     }
 
+    /// Returns the exact size of memory occupied by the texture in VRAM, including mipmaps.
+    std::size_t MemorySize() const {
+        std::size_t size = InnerMemorySize(is_layered);
+        if (is_layered)
+            return size * depth;
+        return size;
+    }
+
+    /// Returns the exact size of the memory occupied by a layer in a texture in VRAM, including
+    /// mipmaps.
+    std::size_t LayerMemorySize() const {
+        return InnerMemorySize(true);
+    }
+
+    /// Returns the size of a layer of this surface in OpenGL.
+    std::size_t LayerSizeGL() const {
+        return SizeInBytesRaw(true) / depth;
+    }
+
     /// Creates SurfaceParams from a texture configuration
     static SurfaceParams CreateForTexture(const Tegra::Texture::FullTextureInfo& config,
                                           const GLShader::SamplerEntry& entry);
@@ -782,6 +818,7 @@ struct SurfaceParams {
     u32 unaligned_height;
     SurfaceTarget target;
     u32 max_mip_level;
+    bool is_layered;
 
     // Parameters used for caching
     VAddr addr;
@@ -797,6 +834,9 @@ struct SurfaceParams {
         u32 layer_stride;
         u32 base_layer;
     } rt;
+
+private:
+    std::size_t InnerMemorySize(bool layer_only = false) const;
 };
 
 }; // namespace OpenGL

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -6,6 +6,7 @@
 #include <set>
 #include <string>
 #include <string_view>
+#include <unordered_set>
 
 #include <boost/optional.hpp>
 #include <fmt/format.h>
@@ -163,10 +164,11 @@ private:
                     const ExitMethod jmp = Scan(target, end, labels);
                     return exit_method = ParallelExit(no_jmp, jmp);
                 }
-                case OpCode::Id::SSY: {
-                    // The SSY instruction uses a similar encoding as the BRA instruction.
+                case OpCode::Id::SSY:
+                case OpCode::Id::PBK: {
+                    // The SSY and PBK use a similar encoding as the BRA instruction.
                     ASSERT_MSG(instr.bra.constant_buffer == 0,
-                               "Constant buffer SSY is not supported");
+                               "Constant buffer branching is not supported");
                     const u32 target = offset + instr.bra.GetBranchTarget();
                     labels.insert(target);
                     // Continue scanning for an exit method.
@@ -275,7 +277,8 @@ public:
     GLSLRegisterManager(ShaderWriter& shader, ShaderWriter& declarations,
                         const Maxwell3D::Regs::ShaderStage& stage, const std::string& suffix,
                         const Tegra::Shader::Header& header)
-        : shader{shader}, declarations{declarations}, stage{stage}, suffix{suffix}, header{header} {
+        : shader{shader}, declarations{declarations}, stage{stage}, suffix{suffix}, header{header},
+          fixed_pipeline_output_attributes_used{}, local_memory_size{0} {
         BuildRegisterList();
         BuildInputList();
     }
@@ -378,8 +381,8 @@ public:
      * @param reg The destination register to use.
      * @param elem The element to use for the operation.
      * @param value The code representing the value to assign. Type has to be half float.
-     * @param type Half float kind of assignment.
-     * @param dest_num_components Number of components in the destionation.
+     * @param merge Half float kind of assignment.
+     * @param dest_num_components Number of components in the destination.
      * @param value_num_components Number of components in the value.
      * @param is_saturated Optional, when True, saturates the provided value.
      * @param dest_elem Optional, the destination element to use for the operation.
@@ -422,6 +425,7 @@ public:
      * @param reg The destination register to use.
      * @param elem The element to use for the operation.
      * @param attribute The input attribute to use as the source value.
+     * @param input_mode The input mode.
      * @param vertex The register that decides which vertex to read from (used in GS).
      */
     void SetRegisterToInputAttibute(const Register& reg, u64 elem, Attribute::Index attribute,
@@ -432,6 +436,25 @@ public:
         shader.AddLine(dest + " = " + src + ';');
     }
 
+    std::string GetLocalMemoryAsFloat(const std::string& index) {
+        return "lmem[" + index + ']';
+    }
+
+    std::string GetLocalMemoryAsInteger(const std::string& index, bool is_signed = false) {
+        const std::string func{is_signed ? "floatToIntBits" : "floatBitsToUint"};
+        return func + "(lmem[" + index + "])";
+    }
+
+    void SetLocalMemoryAsFloat(const std::string& index, const std::string& value) {
+        shader.AddLine("lmem[" + index + "] = " + value + ';');
+    }
+
+    void SetLocalMemoryAsInteger(const std::string& index, const std::string& value,
+                                 bool is_signed = false) {
+        const std::string func{is_signed ? "intBitsToFloat" : "uintBitsToFloat"};
+        shader.AddLine("lmem[" + index + "] = " + func + '(' + value + ");");
+    }
+
     std::string GetControlCode(const Tegra::Shader::ControlCode cc) const {
         switch (cc) {
         case Tegra::Shader::ControlCode::NEU:
@@ -478,7 +501,12 @@ public:
                                std::to_string(static_cast<u32>(attribute)) + ']' +
                                GetSwizzle(elem) + " = " + src + ';');
             } else {
-                shader.AddLine(dest + GetSwizzle(elem) + " = " + src + ';');
+                if (attribute == Attribute::Index::PointSize) {
+                    fixed_pipeline_output_attributes_used.insert(attribute);
+                    shader.AddLine(dest + " = " + src + ';');
+                } else {
+                    shader.AddLine(dest + GetSwizzle(elem) + " = " + src + ';');
+                }
             }
         }
     }
@@ -522,7 +550,9 @@ public:
 
     /// Add declarations.
     void GenerateDeclarations(const std::string& suffix) {
+        GenerateVertex();
         GenerateRegisters(suffix);
+        GenerateLocalMemory();
         GenerateInternalFlags();
         GenerateInputAttrs();
         GenerateOutputAttrs();
@@ -568,6 +598,10 @@ public:
         return entry.GetName();
     }
 
+    void SetLocalMemory(u64 lmem) {
+        local_memory_size = lmem;
+    }
+
 private:
     /// Generates declarations for registers.
     void GenerateRegisters(const std::string& suffix) {
@@ -578,6 +612,15 @@ private:
         declarations.AddNewLine();
     }
 
+    /// Generates declarations for local memory.
+    void GenerateLocalMemory() {
+        if (local_memory_size > 0) {
+            declarations.AddLine("float lmem[" + std::to_string((local_memory_size - 1 + 4) / 4) +
+                                 "];");
+            declarations.AddNewLine();
+        }
+    }
+
     /// Generates declarations for internal flags.
     void GenerateInternalFlags() {
         for (u32 ii = 0; ii < static_cast<u64>(InternalFlag::Amount); ii++) {
@@ -681,6 +724,20 @@ private:
         declarations.AddNewLine();
     }
 
+    void GenerateVertex() {
+        if (stage != Maxwell3D::Regs::ShaderStage::Vertex)
+            return;
+        declarations.AddLine("out gl_PerVertex {");
+        ++declarations.scope;
+        declarations.AddLine("vec4 gl_Position;");
+        for (auto& o : fixed_pipeline_output_attributes_used) {
+            if (o == Attribute::Index::PointSize)
+                declarations.AddLine("float gl_PointSize;");
+        }
+        --declarations.scope;
+        declarations.AddLine("};");
+    }
+
     /// Generates code representing a temporary (GPR) register.
     std::string GetRegister(const Register& reg, unsigned elem) {
         if (reg == Register::ZeroIndex) {
@@ -834,6 +891,8 @@ private:
     /// Generates code representing the declaration name of an output attribute register.
     std::string GetOutputAttribute(Attribute::Index attribute) {
         switch (attribute) {
+        case Attribute::Index::PointSize:
+            return "gl_PointSize";
         case Attribute::Index::Position:
             return "position";
         default:
@@ -868,6 +927,8 @@ private:
     const Maxwell3D::Regs::ShaderStage& stage;
     const std::string& suffix;
     const Tegra::Shader::Header& header;
+    std::unordered_set<Attribute::Index> fixed_pipeline_output_attributes_used;
+    u64 local_memory_size;
 };
 
 class GLSLGenerator {
@@ -877,6 +938,8 @@ public:
         : subroutines(subroutines), program_code(program_code), main_offset(main_offset),
           stage(stage), suffix(suffix) {
         std::memcpy(&header, program_code.data(), sizeof(Tegra::Shader::Header));
+        local_memory_size = header.GetLocalMemorySize();
+        regs.SetLocalMemory(local_memory_size);
         Generate(suffix);
     }
 
@@ -951,7 +1014,7 @@ private:
         // Can't assign to the constant predicate.
         ASSERT(pred != static_cast<u64>(Pred::UnusedIndex));
 
-        const std::string variable = 'p' + std::to_string(pred) + '_' + suffix;
+        std::string variable = 'p' + std::to_string(pred) + '_' + suffix;
         shader.AddLine(variable + " = " + value + ';');
         declr_predicates.insert(std::move(variable));
     }
@@ -1058,7 +1121,7 @@ private:
     /*
      * Transforms the input string GLSL operand into an unpacked half float pair.
      * @note This function returns a float type pair instead of a half float pair. This is because
-     * real half floats are not standarized in GLSL but unpackHalf2x16 (which returns a vec2) is.
+     * real half floats are not standardized in GLSL but unpackHalf2x16 (which returns a vec2) is.
      * @param operand Input operand. It has to be an unsigned integer.
      * @param type How to unpack the unsigned integer to a half float pair.
      * @param abs Get the absolute value of unpacked half floats.
@@ -1232,27 +1295,27 @@ private:
     }
 
     /*
-     * Emits code to push the input target address to the SSY address stack, incrementing the stack
+     * Emits code to push the input target address to the flow address stack, incrementing the stack
      * top.
      */
-    void EmitPushToSSYStack(u32 target) {
+    void EmitPushToFlowStack(u32 target) {
         shader.AddLine('{');
         ++shader.scope;
-        shader.AddLine("ssy_stack[ssy_stack_top] = " + std::to_string(target) + "u;");
-        shader.AddLine("ssy_stack_top++;");
+        shader.AddLine("flow_stack[flow_stack_top] = " + std::to_string(target) + "u;");
+        shader.AddLine("flow_stack_top++;");
         --shader.scope;
         shader.AddLine('}');
     }
 
     /*
-     * Emits code to pop an address from the SSY address stack, setting the jump address to the
+     * Emits code to pop an address from the flow address stack, setting the jump address to the
      * popped address and decrementing the stack top.
      */
-    void EmitPopFromSSYStack() {
+    void EmitPopFromFlowStack() {
         shader.AddLine('{');
         ++shader.scope;
-        shader.AddLine("ssy_stack_top--;");
-        shader.AddLine("jmp_to = ssy_stack[ssy_stack_top];");
+        shader.AddLine("flow_stack_top--;");
+        shader.AddLine("jmp_to = flow_stack[flow_stack_top];");
         shader.AddLine("break;");
         --shader.scope;
         shader.AddLine('}');
@@ -1264,9 +1327,29 @@ private:
 
         ASSERT_MSG(header.ps.omap.sample_mask == 0, "Samplemask write is unimplemented");
 
+        shader.AddLine("if (alpha_test[0] != 0) {");
+        ++shader.scope;
+        // We start on the register containing the alpha value in the first RT.
+        u32 current_reg = 3;
+        for (u32 render_target = 0; render_target < Maxwell3D::Regs::NumRenderTargets;
+             ++render_target) {
+            // TODO(Blinkhawk): verify the behavior of alpha testing on hardware when
+            // multiple render targets are used.
+            if (header.ps.IsColorComponentOutputEnabled(render_target, 0) ||
+                header.ps.IsColorComponentOutputEnabled(render_target, 1) ||
+                header.ps.IsColorComponentOutputEnabled(render_target, 2) ||
+                header.ps.IsColorComponentOutputEnabled(render_target, 3)) {
+                shader.AddLine(fmt::format("if (!AlphaFunc({})) discard;",
+                                           regs.GetRegisterAsFloat(current_reg)));
+                current_reg += 4;
+            }
+        }
+        --shader.scope;
+        shader.AddLine('}');
+
         // Write the color outputs using the data in the shader registers, disabled
         // rendertargets/components are skipped in the register assignment.
-        u32 current_reg = 0;
+        current_reg = 0;
         for (u32 render_target = 0; render_target < Maxwell3D::Regs::NumRenderTargets;
              ++render_target) {
             // TODO(Subv): Figure out how dual-source blending is configured in the Switch.
@@ -1290,6 +1373,63 @@ private:
         }
     }
 
+    /// Unpacks a video instruction operand (e.g. VMAD).
+    std::string GetVideoOperand(const std::string& op, bool is_chunk, bool is_signed,
+                                Tegra::Shader::VideoType type, u64 byte_height) {
+        const std::string value = [&]() {
+            if (!is_chunk) {
+                const auto offset = static_cast<u32>(byte_height * 8);
+                return "((" + op + " >> " + std::to_string(offset) + ") & 0xff)";
+            }
+            const std::string zero = "0";
+
+            switch (type) {
+            case Tegra::Shader::VideoType::Size16_Low:
+                return '(' + op + " & 0xffff)";
+            case Tegra::Shader::VideoType::Size16_High:
+                return '(' + op + " >> 16)";
+            case Tegra::Shader::VideoType::Size32:
+                // TODO(Rodrigo): From my hardware tests it becomes a bit "mad" when
+                // this type is used (1 * 1 + 0 == 0x5b800000). Until a better
+                // explanation is found: assert.
+                UNIMPLEMENTED();
+                return zero;
+            case Tegra::Shader::VideoType::Invalid:
+                UNREACHABLE_MSG("Invalid instruction encoding");
+                return zero;
+            default:
+                UNREACHABLE();
+                return zero;
+            }
+        }();
+
+        if (is_signed) {
+            return "int(" + value + ')';
+        }
+        return value;
+    };
+
+    /// Gets the A operand for a video instruction.
+    std::string GetVideoOperandA(Instruction instr) {
+        return GetVideoOperand(regs.GetRegisterAsInteger(instr.gpr8, 0, false),
+                               instr.video.is_byte_chunk_a != 0, instr.video.signed_a,
+                               instr.video.type_a, instr.video.byte_height_a);
+    }
+
+    /// Gets the B operand for a video instruction.
+    std::string GetVideoOperandB(Instruction instr) {
+        if (instr.video.use_register_b) {
+            return GetVideoOperand(regs.GetRegisterAsInteger(instr.gpr20, 0, false),
+                                   instr.video.is_byte_chunk_b != 0, instr.video.signed_b,
+                                   instr.video.type_b, instr.video.byte_height_b);
+        } else {
+            return '(' +
+                   std::to_string(instr.video.signed_b ? static_cast<s16>(instr.alu.GetImm20_16())
+                                                       : instr.alu.GetImm20_16()) +
+                   ')';
+        }
+    }
+
     /**
      * Compiles a single instruction from Tegra to GLSL.
      * @param offset the offset of the Tegra shader instruction.
@@ -1459,9 +1599,10 @@ private:
                 break;
             }
             case OpCode::Id::FMUL32_IMM: {
-                regs.SetRegisterToFloat(
-                    instr.gpr0, 0,
-                    regs.GetRegisterAsFloat(instr.gpr8) + " * " + GetImmediate32(instr), 1, 1);
+                regs.SetRegisterToFloat(instr.gpr0, 0,
+                                        regs.GetRegisterAsFloat(instr.gpr8) + " * " +
+                                            GetImmediate32(instr),
+                                        1, 1, instr.fmul32.saturate);
                 break;
             }
             case OpCode::Id::FADD32I: {
@@ -2219,6 +2360,39 @@ private:
                 shader.AddLine("}");
                 break;
             }
+            case OpCode::Id::LD_L: {
+                // Add an extra scope and declare the index register inside to prevent
+                // overwriting it in case it is used as an output of the LD instruction.
+                shader.AddLine('{');
+                ++shader.scope;
+
+                std::string op = '(' + regs.GetRegisterAsInteger(instr.gpr8, 0, false) + " + " +
+                                 std::to_string(instr.smem_imm.Value()) + ')';
+
+                shader.AddLine("uint index = (" + op + " / 4);");
+
+                const std::string op_a = regs.GetLocalMemoryAsFloat("index");
+
+                if (instr.ld_l.unknown != 1) {
+                    LOG_CRITICAL(HW_GPU, "LD_L Unhandled mode: {}",
+                                 static_cast<unsigned>(instr.ld_l.unknown.Value()));
+                    UNREACHABLE();
+                }
+
+                switch (instr.ldst_sl.type.Value()) {
+                case Tegra::Shader::StoreType::Bytes32:
+                    regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1);
+                    break;
+                default:
+                    LOG_CRITICAL(HW_GPU, "LD_L Unhandled type: {}",
+                                 static_cast<unsigned>(instr.ldst_sl.type.Value()));
+                    UNREACHABLE();
+                }
+
+                --shader.scope;
+                shader.AddLine('}');
+                break;
+            }
             case OpCode::Id::ST_A: {
                 ASSERT_MSG(instr.gpr8.Value() == Register::ZeroIndex,
                            "Indirect attribute loads are not supported");
@@ -2247,6 +2421,37 @@ private:
 
                 break;
             }
+            case OpCode::Id::ST_L: {
+                // Add an extra scope and declare the index register inside to prevent
+                // overwriting it in case it is used as an output of the LD instruction.
+                shader.AddLine('{');
+                ++shader.scope;
+
+                std::string op = '(' + regs.GetRegisterAsInteger(instr.gpr8, 0, false) + " + " +
+                                 std::to_string(instr.smem_imm.Value()) + ')';
+
+                shader.AddLine("uint index = (" + op + " / 4);");
+
+                if (instr.st_l.unknown != 0) {
+                    LOG_CRITICAL(HW_GPU, "ST_L Unhandled mode: {}",
+                                 static_cast<unsigned>(instr.st_l.unknown.Value()));
+                    UNREACHABLE();
+                }
+
+                switch (instr.ldst_sl.type.Value()) {
+                case Tegra::Shader::StoreType::Bytes32:
+                    regs.SetLocalMemoryAsFloat("index", regs.GetRegisterAsFloat(instr.gpr0));
+                    break;
+                default:
+                    LOG_CRITICAL(HW_GPU, "ST_L Unhandled type: {}",
+                                 static_cast<unsigned>(instr.ldst_sl.type.Value()));
+                    UNREACHABLE();
+                }
+
+                --shader.scope;
+                shader.AddLine('}');
+                break;
+            }
             case OpCode::Id::TEX: {
                 Tegra::Shader::TextureType texture_type{instr.tex.texture_type};
                 std::string coord;
@@ -2736,20 +2941,13 @@ private:
             break;
         }
         case OpCode::Type::FloatSetPredicate: {
-            std::string op_a = instr.fsetp.neg_a ? "-" : "";
-            op_a += regs.GetRegisterAsFloat(instr.gpr8);
+            const std::string op_a =
+                GetOperandAbsNeg(regs.GetRegisterAsFloat(instr.gpr8), instr.fsetp.abs_a != 0,
+                                 instr.fsetp.neg_a != 0);
 
-            if (instr.fsetp.abs_a) {
-                op_a = "abs(" + op_a + ')';
-            }
-
-            std::string op_b{};
+            std::string op_b;
 
             if (instr.is_b_imm) {
-                if (instr.fsetp.neg_b) {
-                    // Only the immediate version of fsetp has a neg_b bit.
-                    op_b += '-';
-                }
                 op_b += '(' + GetImmediate19(instr) + ')';
             } else {
                 if (instr.is_b_gpr) {
@@ -2945,33 +3143,24 @@ private:
             break;
         }
         case OpCode::Type::FloatSet: {
-            std::string op_a = instr.fset.neg_a ? "-" : "";
-            op_a += regs.GetRegisterAsFloat(instr.gpr8);
+            const std::string op_a = GetOperandAbsNeg(regs.GetRegisterAsFloat(instr.gpr8),
+                                                      instr.fset.abs_a != 0, instr.fset.neg_a != 0);
 
-            if (instr.fset.abs_a) {
-                op_a = "abs(" + op_a + ')';
-            }
-
-            std::string op_b = instr.fset.neg_b ? "-" : "";
+            std::string op_b;
 
             if (instr.is_b_imm) {
                 const std::string imm = GetImmediate19(instr);
-                if (instr.fset.neg_imm)
-                    op_b += "(-" + imm + ')';
-                else
-                    op_b += imm;
+                op_b = imm;
             } else {
                 if (instr.is_b_gpr) {
-                    op_b += regs.GetRegisterAsFloat(instr.gpr20);
+                    op_b = regs.GetRegisterAsFloat(instr.gpr20);
                 } else {
-                    op_b += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
-                                            GLSLRegister::Type::Float);
+                    op_b = regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
+                                           GLSLRegister::Type::Float);
                 }
             }
 
-            if (instr.fset.abs_b) {
-                op_b = "abs(" + op_b + ')';
-            }
+            op_b = GetOperandAbsNeg(op_b, instr.fset.abs_b != 0, instr.fset.neg_b != 0);
 
             // The fset instruction sets a register to 1.0 or -1 (depending on the bf bit) if the
             // condition is true, and to 0 otherwise.
@@ -3279,16 +3468,32 @@ private:
                 // The SSY opcode tells the GPU where to re-converge divergent execution paths, it
                 // sets the target of the jump that the SYNC instruction will make. The SSY opcode
                 // has a similar structure to the BRA opcode.
-                ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer SSY is not supported");
+                ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer flow is not supported");
 
                 const u32 target = offset + instr.bra.GetBranchTarget();
-                EmitPushToSSYStack(target);
+                EmitPushToFlowStack(target);
+                break;
+            }
+            case OpCode::Id::PBK: {
+                // PBK pushes to a stack the address where BRK will jump to. This shares stack with
+                // SSY but using SYNC on a PBK address will kill the shader execution. We don't
+                // emulate this because it's very unlikely a driver will emit such invalid shader.
+                ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer PBK is not supported");
+
+                const u32 target = offset + instr.bra.GetBranchTarget();
+                EmitPushToFlowStack(target);
                 break;
             }
             case OpCode::Id::SYNC: {
                 // The SYNC opcode jumps to the address previously set by the SSY opcode
                 ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
-                EmitPopFromSSYStack();
+                EmitPopFromFlowStack();
+                break;
+            }
+            case OpCode::Id::BRK: {
+                // The BRK opcode jumps to the address previously set by the PBK opcode
+                ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
+                EmitPopFromFlowStack();
                 break;
             }
             case OpCode::Id::DEPBAR: {
@@ -3298,87 +3503,51 @@ private:
                 break;
             }
             case OpCode::Id::VMAD: {
-                const bool signed_a = instr.vmad.signed_a == 1;
-                const bool signed_b = instr.vmad.signed_b == 1;
-                const bool result_signed = signed_a || signed_b;
-                boost::optional<std::string> forced_result;
-
-                auto Unpack = [&](const std::string& op, bool is_chunk, bool is_signed,
-                                  Tegra::Shader::VmadType type, u64 byte_height) {
-                    const std::string value = [&]() {
-                        if (!is_chunk) {
-                            const auto offset = static_cast<u32>(byte_height * 8);
-                            return "((" + op + " >> " + std::to_string(offset) + ") & 0xff)";
-                        }
-                        const std::string zero = "0";
-
-                        switch (type) {
-                        case Tegra::Shader::VmadType::Size16_Low:
-                            return '(' + op + " & 0xffff)";
-                        case Tegra::Shader::VmadType::Size16_High:
-                            return '(' + op + " >> 16)";
-                        case Tegra::Shader::VmadType::Size32:
-                            // TODO(Rodrigo): From my hardware tests it becomes a bit "mad" when
-                            // this type is used (1 * 1 + 0 == 0x5b800000). Until a better
-                            // explanation is found: assert.
-                            UNREACHABLE_MSG("Unimplemented");
-                            return zero;
-                        case Tegra::Shader::VmadType::Invalid:
-                            // Note(Rodrigo): This flag is invalid according to nvdisasm. From my
-                            // testing (even though it's invalid) this makes the whole instruction
-                            // assign zero to target register.
-                            forced_result = boost::make_optional(zero);
-                            return zero;
-                        default:
-                            UNREACHABLE();
-                            return zero;
-                        }
-                    }();
-
-                    if (is_signed) {
-                        return "int(" + value + ')';
-                    }
-                    return value;
-                };
-
-                const std::string op_a = Unpack(regs.GetRegisterAsInteger(instr.gpr8, 0, false),
-                                                instr.vmad.is_byte_chunk_a != 0, signed_a,
-                                                instr.vmad.type_a, instr.vmad.byte_height_a);
-
-                std::string op_b;
-                if (instr.vmad.use_register_b) {
-                    op_b = Unpack(regs.GetRegisterAsInteger(instr.gpr20, 0, false),
-                                  instr.vmad.is_byte_chunk_b != 0, signed_b, instr.vmad.type_b,
-                                  instr.vmad.byte_height_b);
-                } else {
-                    op_b = '(' +
-                           std::to_string(signed_b ? static_cast<s16>(instr.alu.GetImm20_16())
-                                                   : instr.alu.GetImm20_16()) +
-                           ')';
-                }
-
+                const bool result_signed = instr.video.signed_a == 1 || instr.video.signed_b == 1;
+                const std::string op_a = GetVideoOperandA(instr);
+                const std::string op_b = GetVideoOperandB(instr);
                 const std::string op_c = regs.GetRegisterAsInteger(instr.gpr39, 0, result_signed);
 
-                std::string result;
-                if (forced_result) {
-                    result = *forced_result;
-                } else {
-                    result = '(' + op_a + " * " + op_b + " + " + op_c + ')';
+                std::string result = '(' + op_a + " * " + op_b + " + " + op_c + ')';
 
-                    switch (instr.vmad.shr) {
-                    case Tegra::Shader::VmadShr::Shr7:
-                        result = '(' + result + " >> 7)";
-                        break;
-                    case Tegra::Shader::VmadShr::Shr15:
-                        result = '(' + result + " >> 15)";
-                        break;
-                    }
+                switch (instr.vmad.shr) {
+                case Tegra::Shader::VmadShr::Shr7:
+                    result = '(' + result + " >> 7)";
+                    break;
+                case Tegra::Shader::VmadShr::Shr15:
+                    result = '(' + result + " >> 15)";
+                    break;
                 }
+
                 regs.SetRegisterToInteger(instr.gpr0, result_signed, 1, result, 1, 1,
                                           instr.vmad.saturate == 1, 0, Register::Size::Word,
                                           instr.vmad.cc);
                 break;
             }
+            case OpCode::Id::VSETP: {
+                const std::string op_a = GetVideoOperandA(instr);
+                const std::string op_b = GetVideoOperandB(instr);
+
+                // We can't use the constant predicate as destination.
+                ASSERT(instr.vsetp.pred3 != static_cast<u64>(Pred::UnusedIndex));
+
+                const std::string second_pred = GetPredicateCondition(instr.vsetp.pred39, false);
+
+                const std::string combiner = GetPredicateCombiner(instr.vsetp.op);
+
+                const std::string predicate = GetPredicateComparison(instr.vsetp.cond, op_a, op_b);
+                // Set the primary predicate to the result of Predicate OP SecondPredicate
+                SetPredicate(instr.vsetp.pred3,
+                             '(' + predicate + ") " + combiner + " (" + second_pred + ')');
+
+                if (instr.vsetp.pred0 != static_cast<u64>(Pred::UnusedIndex)) {
+                    // Set the secondary predicate to the result of !Predicate OP SecondPredicate,
+                    // if enabled
+                    SetPredicate(instr.vsetp.pred0,
+                                 "!(" + predicate + ") " + combiner + " (" + second_pred + ')');
+                }
+                break;
+            }
             default: {
                 LOG_CRITICAL(HW_GPU, "Unhandled instruction: {}", opcode->GetName());
                 UNREACHABLE();
@@ -3442,11 +3611,11 @@ private:
                 labels.insert(subroutine.begin);
                 shader.AddLine("uint jmp_to = " + std::to_string(subroutine.begin) + "u;");
 
-                // TODO(Subv): Figure out the actual depth of the SSY stack, for now it seems
-                // unlikely that shaders will use 20 nested SSYs.
-                constexpr u32 SSY_STACK_SIZE = 20;
-                shader.AddLine("uint ssy_stack[" + std::to_string(SSY_STACK_SIZE) + "];");
-                shader.AddLine("uint ssy_stack_top = 0u;");
+                // TODO(Subv): Figure out the actual depth of the flow stack, for now it seems
+                // unlikely that shaders will use 20 nested SSYs and PBKs.
+                constexpr u32 FLOW_STACK_SIZE = 20;
+                shader.AddLine("uint flow_stack[" + std::to_string(FLOW_STACK_SIZE) + "];");
+                shader.AddLine("uint flow_stack_top = 0u;");
 
                 shader.AddLine("while (true) {");
                 ++shader.scope;
@@ -3506,6 +3675,7 @@ private:
     const u32 main_offset;
     Maxwell3D::Regs::ShaderStage stage;
     const std::string& suffix;
+    u64 local_memory_size;
 
     ShaderWriter shader;
     ShaderWriter declarations;
@@ -3513,7 +3683,7 @@ private:
 
     // Declarations
     std::set<std::string> declr_predicates;
-}; // namespace Decompiler
+}; // namespace OpenGL::GLShader::Decompiler
 
 std::string GetCommonDeclarations() {
     return fmt::format("#define MAX_CONSTBUFFER_ELEMENTS {}\n",

src/video_core/renderer_opengl/gl_shader_gen.cpp

@@ -19,9 +19,6 @@ ProgramResult GenerateVertexShader(const ShaderSetup& setup) {
     out += Decompiler::GetCommonDeclarations();
 
     out += R"(
-out gl_PerVertex {
-    vec4 gl_Position;
-};
 
 layout (location = 0) out vec4 position;
 
@@ -29,6 +26,7 @@ layout(std140) uniform vs_config {
     vec4 viewport_flip;
     uvec4 instance_id;
     uvec4 flip_stage;
+    uvec4 alpha_test;
 };
 )";
 
@@ -105,6 +103,7 @@ layout (std140) uniform gs_config {
     vec4 viewport_flip;
     uvec4 instance_id;
     uvec4 flip_stage;
+    uvec4 alpha_test;
 };
 
 void main() {
@@ -142,8 +141,33 @@ layout (std140) uniform fs_config {
     vec4 viewport_flip;
     uvec4 instance_id;
     uvec4 flip_stage;
+    uvec4 alpha_test;
 };
 
+bool AlphaFunc(in float value) {
+    float ref = uintBitsToFloat(alpha_test[2]);
+    switch (alpha_test[1]) {
+        case 1:
+            return false;
+        case 2:
+            return value < ref;
+        case 3:
+            return value == ref;
+        case 4:
+            return value <= ref;
+        case 5:
+            return value > ref;
+        case 6:
+            return value != ref;
+        case 7:
+            return value >= ref;
+        case 8:
+            return true;
+        default:
+            return false;
+    }
+}
+
 void main() {
     exec_fragment();
 }
@@ -152,4 +176,4 @@ void main() {
     out += program.first;
     return {out, program.second};
 }
-} // namespace OpenGL::GLShader
+} // namespace OpenGL::GLShader

src/video_core/renderer_opengl/gl_shader_manager.cpp

@@ -16,6 +16,17 @@ void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& sh
     viewport_flip[0] = regs.viewport_transform[0].scale_x < 0.0 ? -1.0f : 1.0f;
     viewport_flip[1] = regs.viewport_transform[0].scale_y < 0.0 ? -1.0f : 1.0f;
 
+    u32 func = static_cast<u32>(regs.alpha_test_func);
+    // Normalize the gl variants of opCompare to be the same as the normal variants
+    u32 op_gl_variant_base = static_cast<u32>(Tegra::Engines::Maxwell3D::Regs::ComparisonOp::Never);
+    if (func >= op_gl_variant_base) {
+        func = func - op_gl_variant_base + 1U;
+    }
+
+    alpha_test.enabled = regs.alpha_test_enabled;
+    alpha_test.func = func;
+    alpha_test.ref = regs.alpha_test_ref;
+
     // We only assign the instance to the first component of the vector, the rest is just padding.
     instance_id[0] = state.current_instance;
 

src/video_core/renderer_opengl/gl_shader_manager.h

@@ -22,8 +22,14 @@ struct MaxwellUniformData {
     alignas(16) GLvec4 viewport_flip;
     alignas(16) GLuvec4 instance_id;
     alignas(16) GLuvec4 flip_stage;
+    struct alignas(16) {
+        GLuint enabled;
+        GLuint func;
+        GLfloat ref;
+        GLuint padding;
+    } alpha_test;
 };
-static_assert(sizeof(MaxwellUniformData) == 48, "MaxwellUniformData structure size is incorrect");
+static_assert(sizeof(MaxwellUniformData) == 64, "MaxwellUniformData structure size is incorrect");
 static_assert(sizeof(MaxwellUniformData) < 16384,
               "MaxwellUniformData structure must be less than 16kb as per the OpenGL spec");
 

src/video_core/renderer_opengl/gl_state.cpp

@@ -21,6 +21,8 @@ OpenGLState::OpenGLState() {
     depth.test_enabled = false;
     depth.test_func = GL_LESS;
     depth.write_mask = GL_TRUE;
+    depth.depth_range_near = 0.0f;
+    depth.depth_range_far = 1.0f;
 
     color_mask.red_enabled = GL_TRUE;
     color_mask.green_enabled = GL_TRUE;
@@ -119,6 +121,12 @@ void OpenGLState::Apply() const {
         glDepthMask(depth.write_mask);
     }
 
+    // Depth range
+    if (depth.depth_range_near != cur_state.depth.depth_range_near ||
+        depth.depth_range_far != cur_state.depth.depth_range_far) {
+        glDepthRange(depth.depth_range_near, depth.depth_range_far);
+    }
+
     // Color mask
     if (color_mask.red_enabled != cur_state.color_mask.red_enabled ||
         color_mask.green_enabled != cur_state.color_mask.green_enabled ||

src/video_core/renderer_opengl/gl_state.h

@@ -42,9 +42,11 @@ public:
     } cull;
 
     struct {
-        bool test_enabled;    // GL_DEPTH_TEST
-        GLenum test_func;     // GL_DEPTH_FUNC
-        GLboolean write_mask; // GL_DEPTH_WRITEMASK
+        bool test_enabled;        // GL_DEPTH_TEST
+        GLenum test_func;         // GL_DEPTH_FUNC
+        GLboolean write_mask;     // GL_DEPTH_WRITEMASK
+        GLfloat depth_range_near; // GL_DEPTH_RANGE
+        GLfloat depth_range_far;  // GL_DEPTH_RANGE
     } depth;
 
     struct {

src/video_core/renderer_opengl/maxwell_to_gl.h

@@ -82,8 +82,20 @@ inline GLenum VertexType(Maxwell::VertexAttribute attrib) {
         return {};
     }
 
-    case Maxwell::VertexAttribute::Type::Float:
-        return GL_FLOAT;
+    case Maxwell::VertexAttribute::Type::Float: {
+        switch (attrib.size) {
+        case Maxwell::VertexAttribute::Size::Size_16:
+        case Maxwell::VertexAttribute::Size::Size_16_16:
+        case Maxwell::VertexAttribute::Size::Size_16_16_16:
+        case Maxwell::VertexAttribute::Size::Size_16_16_16_16:
+            return GL_HALF_FLOAT;
+        case Maxwell::VertexAttribute::Size::Size_32:
+        case Maxwell::VertexAttribute::Size::Size_32_32:
+        case Maxwell::VertexAttribute::Size::Size_32_32_32:
+        case Maxwell::VertexAttribute::Size::Size_32_32_32_32:
+            return GL_FLOAT;
+        }
+    }
     }
 
     LOG_CRITICAL(Render_OpenGL, "Unimplemented vertex type={}", attrib.TypeString());

src/video_core/textures/decoders.cpp

@@ -142,7 +142,6 @@ void SwizzledData(u8* swizzled_data, u8* unswizzled_data, const bool unswizzle,
     const u32 blocks_on_x = div_ceil(width, block_x_elements);
     const u32 blocks_on_y = div_ceil(height, block_y_elements);
     const u32 blocks_on_z = div_ceil(depth, block_z_elements);
-    const u32 blocks = blocks_on_x * blocks_on_y * blocks_on_z;
     const u32 gob_size = gob_x_bytes * gob_elements_y * gob_elements_z;
     const u32 xy_block_size = gob_size * block_height;
     const u32 block_size = xy_block_size * block_depth;
@@ -320,13 +319,13 @@ std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat
 std::size_t CalculateSize(bool tiled, u32 bytes_per_pixel, u32 width, u32 height, u32 depth,
                           u32 block_height, u32 block_depth) {
     if (tiled) {
-        const u32 gobs_in_x = 64 / bytes_per_pixel;
+        const u32 gobs_in_x = 64;
         const u32 gobs_in_y = 8;
         const u32 gobs_in_z = 1;
-        const u32 aligned_width = Common::AlignUp(width, gobs_in_x);
+        const u32 aligned_width = Common::AlignUp(width * bytes_per_pixel, gobs_in_x);
         const u32 aligned_height = Common::AlignUp(height, gobs_in_y * block_height);
         const u32 aligned_depth = Common::AlignUp(depth, gobs_in_z * block_depth);
-        return aligned_width * aligned_height * aligned_depth * bytes_per_pixel;
+        return aligned_width * aligned_height * aligned_depth;
     } else {
         return width * height * depth * bytes_per_pixel;
     }

src/web_service/CMakeLists.txt

@@ -10,7 +10,7 @@ add_library(web_service STATIC
 create_target_directory_groups(web_service)
 
 get_directory_property(OPENSSL_LIBS
-        DIRECTORY ${CMAKE_SOURCE_DIR}/externals/libressl
+        DIRECTORY ${PROJECT_SOURCE_DIR}/externals/libressl
         DEFINITION OPENSSL_LIBS)
-target_compile_definitions(web_service PUBLIC -DCPPHTTPLIB_OPENSSL_SUPPORT)
+target_compile_definitions(web_service PRIVATE -DCPPHTTPLIB_OPENSSL_SUPPORT)
 target_link_libraries(web_service PRIVATE common json-headers ${OPENSSL_LIBS} httplib lurlparser)

src/yuzu/CMakeLists.txt

@@ -82,10 +82,10 @@ set(UIS
 )
 
 file(GLOB COMPAT_LIST
-     ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.qrc
-     ${CMAKE_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
-file(GLOB_RECURSE ICONS ${CMAKE_SOURCE_DIR}/dist/icons/*)
-file(GLOB_RECURSE THEMES ${CMAKE_SOURCE_DIR}/dist/qt_themes/*)
+     ${PROJECT_BINARY_DIR}/dist/compatibility_list/compatibility_list.qrc
+     ${PROJECT_BINARY_DIR}/dist/compatibility_list/compatibility_list.json)
+file(GLOB_RECURSE ICONS ${PROJECT_SOURCE_DIR}/dist/icons/*)
+file(GLOB_RECURSE THEMES ${PROJECT_SOURCE_DIR}/dist/qt_themes/*)
 
 qt5_wrap_ui(UI_HDRS ${UIS})
 
@@ -121,7 +121,7 @@ target_link_libraries(yuzu PRIVATE Boost::boost glad Qt5::OpenGL Qt5::Widgets)
 target_link_libraries(yuzu PRIVATE ${PLATFORM_LIBRARIES} Threads::Threads)
 
 if (YUZU_ENABLE_COMPATIBILITY_REPORTING)
-    add_definitions(-DYUZU_ENABLE_COMPATIBILITY_REPORTING)
+    target_compile_definitions(yuzu PRIVATE -DYUZU_ENABLE_COMPATIBILITY_REPORTING)
 endif()
 
 if (USE_DISCORD_PRESENCE)

src/yuzu/bootmanager.cpp

@@ -8,7 +8,6 @@
 
 #include "common/microprofile.h"
 #include "common/scm_rev.h"
-#include "common/string_util.h"
 #include "core/core.h"
 #include "core/frontend/framebuffer_layout.h"
 #include "core/settings.h"
@@ -107,9 +106,8 @@ private:
 GRenderWindow::GRenderWindow(QWidget* parent, EmuThread* emu_thread)
     : QWidget(parent), child(nullptr), emu_thread(emu_thread) {
 
-    std::string window_title = fmt::format("yuzu {} | {}-{}", Common::g_build_name,
-                                           Common::g_scm_branch, Common::g_scm_desc);
-    setWindowTitle(QString::fromStdString(window_title));
+    setWindowTitle(QStringLiteral("yuzu %1 | %2-%3")
+                       .arg(Common::g_build_name, Common::g_scm_branch, Common::g_scm_desc));
     setAttribute(Qt::WA_AcceptTouchEvents);
 
     InputCommon::Init();

src/yuzu/configuration/config.cpp

@@ -4,6 +4,7 @@
 
 #include <QSettings>
 #include "common/file_util.h"
+#include "core/hle/service/acc/profile_manager.h"
 #include "input_common/main.h"
 #include "yuzu/configuration/config.h"
 #include "yuzu/ui_settings.h"
@@ -12,11 +13,16 @@ Config::Config() {
     // TODO: Don't hardcode the path; let the frontend decide where to put the config files.
     qt_config_loc = FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) + "qt-config.ini";
     FileUtil::CreateFullPath(qt_config_loc);
-    qt_config = new QSettings(QString::fromStdString(qt_config_loc), QSettings::IniFormat);
+    qt_config =
+        std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), QSettings::IniFormat);
 
     Reload();
 }
 
+Config::~Config() {
+    Save();
+}
+
 const std::array<int, Settings::NativeButton::NumButtons> Config::default_buttons = {
     Qt::Key_A, Qt::Key_S, Qt::Key_Z,    Qt::Key_X,  Qt::Key_3,     Qt::Key_4,    Qt::Key_Q,
     Qt::Key_W, Qt::Key_1, Qt::Key_2,    Qt::Key_N,  Qt::Key_M,     Qt::Key_F,    Qt::Key_T,
@@ -122,7 +128,11 @@ void Config::ReadValues() {
 
     qt_config->beginGroup("System");
     Settings::values.use_docked_mode = qt_config->value("use_docked_mode", false).toBool();
-    Settings::values.username = qt_config->value("username", "yuzu").toString().toStdString();
+    Settings::values.enable_nfc = qt_config->value("enable_nfc", true).toBool();
+
+    Settings::values.current_user = std::clamp<int>(qt_config->value("current_user", 0).toInt(), 0,
+                                                    Service::Account::MAX_USERS - 1);
+
     Settings::values.language_index = qt_config->value("language_index", 1).toInt();
     qt_config->endGroup();
 
@@ -258,7 +268,9 @@ void Config::SaveValues() {
 
     qt_config->beginGroup("System");
     qt_config->setValue("use_docked_mode", Settings::values.use_docked_mode);
-    qt_config->setValue("username", QString::fromStdString(Settings::values.username));
+    qt_config->setValue("enable_nfc", Settings::values.enable_nfc);
+    qt_config->setValue("current_user", Settings::values.current_user);
+
     qt_config->setValue("language_index", Settings::values.language_index);
     qt_config->endGroup();
 
@@ -335,9 +347,3 @@ void Config::Reload() {
 void Config::Save() {
     SaveValues();
 }
-
-Config::~Config() {
-    Save();
-
-    delete qt_config;
-}