154 files changed, 3632 insertions, 832 deletions
diff --git a/.gitattributes b/.gitattributes
new file mode 100644
index 000000000..ab861a396
--- /dev/null
+++ b/.gitattributes
@@ -0,0 +1,4 @@
+dist/languages/* linguist-vendored
+dist/qt_themes/* linguist-vendored
+externals/* linguist-vendored
+*.h linguist-language=cpp
diff --git a/.travis.yml b/.travis.yml
index dee34a8e3..b0fbe3c5f 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -24,11 +24,24 @@ matrix:
    - os: osx
      env: NAME="macos build"
      sudo: false
-      osx_image: xcode9.3
+      osx_image: xcode10
      install: "./.travis/macos/deps.sh"
      script: "./.travis/macos/build.sh"
      after_success: "./.travis/macos/upload.sh"
      cache: ccache
+    - os: linux
+      env: NAME="MinGW build"
+      sudo: required
+      dist: trusty
+      services: docker
+      addons:
+        apt:
+          packages:
+            - p7zip-full
+      install: "./.travis/linux-mingw/deps.sh"
+      script: "./.travis/linux-mingw/build.sh"
+      after_success: "./.travis/linux-mingw/upload.sh"
+      cache: ccache
 deploy:
  provider: releases
diff --git a/.travis/common/post-upload.sh b/.travis/common/post-upload.sh
index 90deaaec8..28735a9cf 100755
--- a/.travis/common/post-upload.sh
+++ b/.travis/common/post-upload.sh
@@ -11,6 +11,9 @@ if [ -z $TRAVIS_TAG ]; then
    RELEASE_NAME=head
 else
    RELEASE_NAME=$(echo $TRAVIS_TAG | cut -d- -f1)
+    if [ "$NAME" = "MinGW build" ]; then
+        RELEASE_NAME="${RELEASE_NAME}-mingw"
+    fi
 fi
 mv "$REV_NAME" $RELEASE_NAME
diff --git a/.travis/linux-mingw/build.sh b/.travis/linux-mingw/build.sh
new file mode 100755
index 000000000..be03cc0f3
--- /dev/null
+++ b/.travis/linux-mingw/build.sh
@@ -0,0 +1,3 @@
+#!/bin/bash -ex
+mkdir "$HOME/.ccache" || true
+docker run --env-file .travis/common/travis-ci.env -v $(pwd):/yuzu -v "$HOME/.ccache":/root/.ccache ubuntu:18.04 /bin/bash -ex /yuzu/.travis/linux-mingw/docker.sh
diff --git a/.travis/linux-mingw/deps.sh b/.travis/linux-mingw/deps.sh
new file mode 100755
index 000000000..540bb934a
--- /dev/null
+++ b/.travis/linux-mingw/deps.sh
@@ -0,0 +1,3 @@
+#!/bin/sh -ex
+docker pull ubuntu:18.04
diff --git a/.travis/linux-mingw/docker.sh b/.travis/linux-mingw/docker.sh
new file mode 100755
index 000000000..d15c3f6e8
--- /dev/null
+++ b/.travis/linux-mingw/docker.sh
@@ -0,0 +1,59 @@
+#!/bin/bash -ex
+cd /yuzu
+MINGW_PACKAGES="sdl2-mingw-w64 qt5base-mingw-w64 qt5tools-mingw-w64 libsamplerate-mingw-w64 qt5multimedia-mingw-w64"
+apt-get update
+apt-get install -y gpg wget git python3-pip python ccache g++-mingw-w64-x86-64 gcc-mingw-w64-x86-64 mingw-w64-tools cmake
+echo 'deb http://ppa.launchpad.net/tobydox/mingw-w64/ubuntu bionic main ' > /etc/apt/sources.list.d/extras.list
+apt-key adv --keyserver keyserver.ubuntu.com --recv '72931B477E22FEFD47F8DECE02FE5F12ADDE29B2'
+apt-get update
+apt-get install -y ${MINGW_PACKAGES}
+# fix a problem in current MinGW headers
+wget -q https://raw.githubusercontent.com/Alexpux/mingw-w64/d0d7f784833bbb0b2d279310ddc6afb52fe47a46/mingw-w64-headers/crt/errno.h -O /usr/x86_64-w64-mingw32/include/errno.h
+# override Travis CI unreasonable ccache size
+echo 'max_size = 3.0G' > "$HOME/.ccache/ccache.conf"
+# Dirty hack to trick unicorn makefile into believing we are in a MINGW system
+mv /bin/uname /bin/uname1 && echo -e '#!/bin/sh\necho MINGW64' >> /bin/uname
+chmod +x /bin/uname
+# Dirty hack to trick unicorn makefile into believing we have cmd
+echo '' >> /bin/cmd
+chmod +x /bin/cmd
+mkdir build && cd build
+cmake .. -DCMAKE_TOOLCHAIN_FILE="$(pwd)/../CMakeModules/MinGWCross.cmake" -DUSE_CCACHE=ON -DYUZU_USE_BUNDLED_UNICORN=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DCMAKE_BUILD_TYPE=Release
+make -j4
+# Clean up the dirty hacks
+rm /bin/uname && mv /bin/uname1 /bin/uname
+rm /bin/cmd
+ccache -s
+echo "Tests skipped"
+#ctest -VV -C Release
+echo 'Prepare binaries...'
+cd ..
+mkdir package
+QT_PLATFORM_DLL_PATH='/usr/x86_64-w64-mingw32/lib/qt5/plugins/platforms/'
+find build/ -name "yuzu*.exe" -exec cp {} 'package' \;
+# copy Qt plugins
+mkdir package/platforms
+cp "${QT_PLATFORM_DLL_PATH}/qwindows.dll" package/platforms/
+cp -rv "${QT_PLATFORM_DLL_PATH}/../mediaservice/" package/
+cp -rv "${QT_PLATFORM_DLL_PATH}/../imageformats/" package/
+rm -f package/mediaservice/*d.dll
+for i in package/*.exe; do
+  # we need to process pdb here, however, cv2pdb
+  # does not work here, so we just simply strip all the debug symbols
+  x86_64-w64-mingw32-strip "${i}"
+done
+pip3 install pefile
+python3 .travis/linux-mingw/scan_dll.py package/*.exe "package/"
diff --git a/.travis/linux-mingw/scan_dll.py b/.travis/linux-mingw/scan_dll.py
new file mode 100644
index 000000000..163183f2e
--- /dev/null
+++ b/.travis/linux-mingw/scan_dll.py
@@ -0,0 +1,106 @@
+import pefile
+import sys
+import re
+import os
+import queue
+import shutil
+# constant definitions
+KNOWN_SYS_DLLS = ['WINMM.DLL', 'MSVCRT.DLL', 'VERSION.DLL', 'MPR.DLL',
+                  'DWMAPI.DLL', 'UXTHEME.DLL', 'DNSAPI.DLL', 'IPHLPAPI.DLL']
+# below is for Ubuntu 18.04 with specified PPA enabled, if you are using
+# other distro or different repositories, change the following accordingly
+DLL_PATH = [
+    '/usr/x86_64-w64-mingw32/bin/',
+    '/usr/x86_64-w64-mingw32/lib/',
+    '/usr/lib/gcc/x86_64-w64-mingw32/7.3-posix/'
+]
+missing = []
+def parse_imports(file_name):
+    results = []
+    pe = pefile.PE(file_name, fast_load=True)
+    pe.parse_data_directories()
+    for entry in pe.DIRECTORY_ENTRY_IMPORT:
+        current = entry.dll.decode()
+        current_u = current.upper()  # b/c Windows is often case insensitive
+        # here we filter out system dlls
+        # dll w/ names like *32.dll are likely to be system dlls
+        if current_u.upper() not in KNOWN_SYS_DLLS and not re.match(string=current_u, pattern=r'.*32\.DLL'):
+            results.append(current)
+    return results
+def parse_imports_recursive(file_name, path_list=[]):
+    q = queue.Queue()  # create a FIFO queue
+    # file_name can be a string or a list for the convience
+    if isinstance(file_name, str):
+        q.put(file_name)
+    elif isinstance(file_name, list):
+        for i in file_name:
+            q.put(i)
+    full_list = []
+    while q.qsize():
+        current = q.get_nowait()
+        print('> %s' % current)
+        deps = parse_imports(current)
+        # if this dll does not have any import, ignore it
+        if not deps:
+            continue
+        for dep in deps:
+            # the dependency already included in the list, skip
+            if dep in full_list:
+                continue
+            # find the requested dll in the provided paths
+            full_path = find_dll(dep)
+            if not full_path:
+                missing.append(dep)
+                continue
+            full_list.append(dep)
+            q.put(full_path)
+            path_list.append(full_path)
+    return full_list
+def find_dll(name):
+    for path in DLL_PATH:
+        for root, _, files in os.walk(path):
+            for f in files:
+                if name.lower() == f.lower():
+                    return os.path.join(root, f)
+def deploy(name, dst, dry_run=False):
+    dlls_path = []
+    parse_imports_recursive(name, dlls_path)
+    for dll_entry in dlls_path:
+        if not dry_run:
+            shutil.copy(dll_entry, dst)
+        else:
+            print('[Dry-Run] Copy %s to %s' % (dll_entry, dst))
+    print('Deploy completed.')
+    return dlls_path
+def main():
+    if len(sys.argv) < 3:
+        print('Usage: %s [files to examine ...] [target deploy directory]')
+        return 1
+    to_deploy = sys.argv[1:-1]
+    tgt_dir = sys.argv[-1]
+    if not os.path.isdir(tgt_dir):
+        print('%s is not a directory.' % tgt_dir)
+        return 1
+    print('Scanning dependencies...')
+    deploy(to_deploy, tgt_dir)
+    if missing:
+        print('Following DLLs are not found: %s' % ('\n'.join(missing)))
+    return 0
+if __name__ == '__main__':
+    main()
diff --git a/.travis/linux-mingw/upload.sh b/.travis/linux-mingw/upload.sh
new file mode 100755
index 000000000..66e896bc4
--- /dev/null
+++ b/.travis/linux-mingw/upload.sh
@@ -0,0 +1,13 @@
+#!/bin/bash -ex
+. .travis/common/pre-upload.sh
+REV_NAME="yuzu-windows-mingw-${GITDATE}-${GITREV}"
+ARCHIVE_NAME="${REV_NAME}.tar.gz"
+COMPRESSION_FLAGS="-czvf"
+mkdir "$REV_NAME"
+# get around the permission issues
+cp -r package/* "$REV_NAME"
+. .travis/common/post-upload.sh
diff --git a/.travis/linux/docker.sh b/.travis/linux/docker.sh
index 459d6bc75..892d2480a 100755
--- a/.travis/linux/docker.sh
+++ b/.travis/linux/docker.sh
@@ -6,7 +6,9 @@ apt-get install --no-install-recommends -y build-essential git libqt5opengl5-dev
 cd /yuzu
 mkdir build && cd build
-cmake .. -DYUZU_BUILD_UNICORN=ON -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=/usr/lib/ccache/gcc -DCMAKE_CXX_COMPILER=/usr/lib/ccache/g++ -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -G Ninja
+cmake .. -DYUZU_USE_BUNDLED_UNICORN=ON -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=/usr/lib/ccache/gcc -DCMAKE_CXX_COMPILER=/usr/lib/ccache/g++ -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -G Ninja
 ninja
+ccache -s
 ctest -VV -C Release
diff --git a/.travis/macos/build.sh b/.travis/macos/build.sh
index b76a153be..e68dc1400 100755
--- a/.travis/macos/build.sh
+++ b/.travis/macos/build.sh
@@ -2,14 +2,16 @@
 set -o pipefail
-export MACOSX_DEPLOYMENT_TARGET=10.12
+export MACOSX_DEPLOYMENT_TARGET=10.13
 export Qt5_DIR=$(brew --prefix)/opt/qt5
 export UNICORNDIR=$(pwd)/externals/unicorn
 export PATH="/usr/local/opt/ccache/libexec:$PATH"
 mkdir build && cd build
 cmake --version
-cmake .. -DYUZU_BUILD_UNICORN=ON -DCMAKE_BUILD_TYPE=Release -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON
+cmake .. -DYUZU_USE_BUNDLED_UNICORN=ON -DCMAKE_BUILD_TYPE=Release -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON
 make -j4
+ccache -s
 ctest -VV -C Release
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 500d099fc..cd990188e 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -123,8 +123,6 @@ else()
    # Avoid windows.h from including some usually unused libs like winsocks.h, since this might cause some redefinition errors.
    add_definitions(/DWIN32_LEAN_AND_MEAN)
-    # set up output paths for executable binaries (.exe-files, and .dll-files on DLL-capable platforms)
-    set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)
    set(CMAKE_CONFIGURATION_TYPES Debug Release CACHE STRING "" FORCE)
    # Tweak optimization settings
@@ -269,10 +267,18 @@ if (YUZU_USE_BUNDLED_UNICORN)
        find_package(PythonInterp 2.7 REQUIRED)
-        add_custom_command(OUTPUT ${LIBUNICORN_LIBRARY}
+        if (MINGW)
-            COMMAND ${CMAKE_COMMAND} -E env UNICORN_ARCHS="aarch64" PYTHON="${PYTHON_EXECUTABLE}" /bin/sh make.sh macos-universal-no
+            add_custom_command(OUTPUT ${LIBUNICORN_LIBRARY}
-            WORKING_DIRECTORY ${UNICORN_PREFIX}
+                COMMAND ${CMAKE_COMMAND} -E env UNICORN_ARCHS="aarch64" PYTHON="${PYTHON_EXECUTABLE}" /bin/sh make.sh cross-win64
-        )
+                WORKING_DIRECTORY ${UNICORN_PREFIX}
+            )
+        else()
+            add_custom_command(OUTPUT ${LIBUNICORN_LIBRARY}
+                COMMAND ${CMAKE_COMMAND} -E env UNICORN_ARCHS="aarch64" PYTHON="${PYTHON_EXECUTABLE}" /bin/sh make.sh macos-universal-no
+                WORKING_DIRECTORY ${UNICORN_PREFIX}
+            )
+        endif()
        # ALL makes this custom target build every time
        # but it won't actually build if LIBUNICORN_LIBRARY is up to date
        add_custom_target(unicorn-build ALL
@@ -286,6 +292,7 @@ endif()
 if (UNICORN_FOUND)
    add_library(unicorn INTERFACE)
+    add_dependencies(unicorn unicorn-build)
    target_link_libraries(unicorn INTERFACE "${LIBUNICORN_LIBRARY}")
    target_include_directories(unicorn INTERFACE "${LIBUNICORN_INCLUDE_DIR}")
 else()
@@ -431,8 +438,12 @@ enable_testing()
 add_subdirectory(externals)
 add_subdirectory(src)
-# Set yuzu project as default StartUp Project in Visual Studio
+# Set yuzu project or yuzu-cmd project as default StartUp Project in Visual Studio depending on whether QT is enabled or not
-set_property(DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR} PROPERTY VS_STARTUP_PROJECT yuzu)
+if(ENABLE_QT)
+    set_property(DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR} PROPERTY VS_STARTUP_PROJECT yuzu)
+else()
+    set_property(DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR} PROPERTY VS_STARTUP_PROJECT yuzu-cmd)
+endif()
 # Installation instructions
diff --git a/CMakeModules/MinGWCross.cmake b/CMakeModules/MinGWCross.cmake
new file mode 100644
index 000000000..29ecd1ac4
--- /dev/null
+++ b/CMakeModules/MinGWCross.cmake
@@ -0,0 +1,54 @@
+set(MINGW_PREFIX   /usr/x86_64-w64-mingw32/)
+set(CMAKE_SYSTEM_NAME               Windows)
+set(CMAKE_SYSTEM_PROCESSOR           x86_64)
+# Actually a hack, w/o this will cause some strange errors
+set(CMAKE_HOST_WIN32                 TRUE)
+set(CMAKE_FIND_ROOT_PATH            ${MINGW_PREFIX})
+set(SDL2_PATH                       ${MINGW_PREFIX})
+set(MINGW_TOOL_PREFIX               ${CMAKE_SYSTEM_PROCESSOR}-w64-mingw32-)
+# Specify the cross compiler
+set(CMAKE_C_COMPILER            ${MINGW_TOOL_PREFIX}gcc-posix)
+set(CMAKE_CXX_COMPILER          ${MINGW_TOOL_PREFIX}g++-posix)
+set(CMAKE_RC_COMPILER           ${MINGW_TOOL_PREFIX}windres)
+# Mingw tools
+set(STRIP                       ${MINGW_TOOL_PREFIX}strip)
+set(WINDRES                     ${MINGW_TOOL_PREFIX}windres)
+set(ENV{PKG_CONFIG}             ${MINGW_TOOL_PREFIX}pkg-config)
+# ccache wrapper
+option(USE_CCACHE "Use ccache for compilation" OFF)
+if(USE_CCACHE)
+    find_program(CCACHE ccache)
+    if(CCACHE)
+        message(STATUS "Using ccache found in PATH")
+        set_property(GLOBAL PROPERTY RULE_LAUNCH_COMPILE ${CCACHE})
+        set_property(GLOBAL PROPERTY RULE_LAUNCH_LINK ${CCACHE})
+    else(CCACHE)
+        message(WARNING "USE_CCACHE enabled, but no ccache found")
+    endif(CCACHE)
+endif(USE_CCACHE)
+# Search for programs in the build host directories
+set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER)
+# Echo modified cmake vars to screen for debugging purposes
+if(NOT DEFINED ENV{MINGW_DEBUG_INFO})
+        message("")
+        message("Custom cmake vars: (blank = system default)")
+        message("-----------------------------------------")
+        message("* CMAKE_C_COMPILER                     : ${CMAKE_C_COMPILER}")
+        message("* CMAKE_CXX_COMPILER                   : ${CMAKE_CXX_COMPILER}")
+        message("* CMAKE_RC_COMPILER                    : ${CMAKE_RC_COMPILER}")
+        message("* WINDRES                              : ${WINDRES}")
+        message("* ENV{PKG_CONFIG}                      : $ENV{PKG_CONFIG}")
+        message("* STRIP                                : ${STRIP}")
+        message("* USE_CCACHE                           : ${USE_CCACHE}")
+        message("")
+        # So that the debug info only appears once
+        set(ENV{MINGW_DEBUG_INFO} SHOWN)
+endif()
diff --git a/externals/dynarmic b/externals/dynarmic
-Subproject 959446573f3adfcba173ef4b0011a4a280f18eb
+Subproject 4e6848d1c9e8dadc70595c15b5589f8b14aad47
diff --git a/src/audio_core/audio_renderer.cpp b/src/audio_core/audio_renderer.cpp
index 83b75e61f..6f0ff953a 100644
--- a/src/audio_core/audio_renderer.cpp
+++ b/src/audio_core/audio_renderer.cpp
@@ -79,6 +79,10 @@ u32 AudioRenderer::GetMixBufferCount() const {
    return worker_params.mix_buffer_count;
 }
+Stream::State AudioRenderer::GetStreamState() const {
+    return stream->GetState();
+}
 std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_params) {
    // Copy UpdateDataHeader struct
    UpdateDataHeader config{};
diff --git a/src/audio_core/audio_renderer.h b/src/audio_core/audio_renderer.h
index 2c4f5ab75..dfef89e1d 100644
--- a/src/audio_core/audio_renderer.h
+++ b/src/audio_core/audio_renderer.h
@@ -170,6 +170,7 @@ public:
    u32 GetSampleRate() const;
    u32 GetSampleCount() const;
    u32 GetMixBufferCount() const;
+    Stream::State GetStreamState() const;
 private:
    class VoiceState;
diff --git a/src/audio_core/stream.cpp b/src/audio_core/stream.cpp
index 449db2416..742a5e0a0 100644
--- a/src/audio_core/stream.cpp
+++ b/src/audio_core/stream.cpp
@@ -49,9 +49,14 @@ void Stream::Play() {
 }
 void Stream::Stop() {
+    state = State::Stopped;
    ASSERT_MSG(false, "Unimplemented");
 }
+Stream::State Stream::GetState() const {
+    return state;
+}
 s64 Stream::GetBufferReleaseCycles(const Buffer& buffer) const {
    const std::size_t num_samples{buffer.GetSamples().size() / GetNumChannels()};
    return CoreTiming::usToCycles((static_cast<u64>(num_samples) * 1000000) / sample_rate);
diff --git a/src/audio_core/stream.h b/src/audio_core/stream.h
index 27db1112f..aebfeb51d 100644
--- a/src/audio_core/stream.h
+++ b/src/audio_core/stream.h
@@ -33,6 +33,12 @@ public:
        Multi51Channel16,
    };
+    /// Current state of the stream
+    enum class State {
+        Stopped,
+        Playing,
+    };
    /// Callback function type, used to change guest state on a buffer being released
    using ReleaseCallback = std::function<void()>;
@@ -72,13 +78,10 @@ public:
    /// Gets the number of channels
    u32 GetNumChannels() const;
-private:
+    /// Get the state
-    /// Current state of the stream
+    State GetState() const;
-    enum class State {
-        Stopped,
-        Playing,
-    };
+private:
    /// Plays the next queued buffer in the audio stream, starting playback if necessary
    void PlayNextBuffer();
diff --git a/src/audio_core/time_stretch.cpp b/src/audio_core/time_stretch.cpp
index fc14151da..d72d67994 100644
--- a/src/audio_core/time_stretch.cpp
+++ b/src/audio_core/time_stretch.cpp
@@ -59,7 +59,7 @@ std::size_t TimeStretcher::Process(const s16* in, std::size_t num_in, s16* out,
    m_stretch_ratio = std::max(m_stretch_ratio, 0.05);
    m_sound_touch.setTempo(m_stretch_ratio);
-    LOG_DEBUG(Audio, "{:5}/{:5} ratio:{:0.6f} backlog:{:0.6f}", num_in, num_out, m_stretch_ratio,
+    LOG_TRACE(Audio, "{:5}/{:5} ratio:{:0.6f} backlog:{:0.6f}", num_in, num_out, m_stretch_ratio,
              backlog_fullness);
    m_sound_touch.putSamples(in, static_cast<u32>(num_in));
diff --git a/src/common/common_paths.h b/src/common/common_paths.h
index df2ce80b1..4f88de768 100644
--- a/src/common/common_paths.h
+++ b/src/common/common_paths.h
@@ -33,6 +33,8 @@
 #define NAND_DIR "nand"
 #define SYSDATA_DIR "sysdata"
 #define KEYS_DIR "keys"
+#define LOAD_DIR "load"
+#define DUMP_DIR "dump"
 #define LOG_DIR "log"
 // Filenames
diff --git a/src/common/file_util.cpp b/src/common/file_util.cpp
index 21a0b9738..548463787 100644
--- a/src/common/file_util.cpp
+++ b/src/common/file_util.cpp
@@ -705,6 +705,8 @@ const std::string& GetUserPath(UserPath path, const std::string& new_path) {
 #endif
        paths.emplace(UserPath::SDMCDir, user_path + SDMC_DIR DIR_SEP);
        paths.emplace(UserPath::NANDDir, user_path + NAND_DIR DIR_SEP);
+        paths.emplace(UserPath::LoadDir, user_path + LOAD_DIR DIR_SEP);
+        paths.emplace(UserPath::DumpDir, user_path + DUMP_DIR DIR_SEP);
        paths.emplace(UserPath::SysDataDir, user_path + SYSDATA_DIR DIR_SEP);
        paths.emplace(UserPath::KeysDir, user_path + KEYS_DIR DIR_SEP);
        // TODO: Put the logs in a better location for each OS
diff --git a/src/common/file_util.h b/src/common/file_util.h
index 24c1e413c..3d8fe6264 100644
--- a/src/common/file_util.h
+++ b/src/common/file_util.h
@@ -29,6 +29,8 @@ enum class UserPath {
    NANDDir,
    RootDir,
    SDMCDir,
+    LoadDir,
+    DumpDir,
    SysDataDir,
    UserDir,
 };
diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp
index efd776db6..9f5918851 100644
--- a/src/common/logging/backend.cpp
+++ b/src/common/logging/backend.cpp
@@ -183,6 +183,7 @@ void FileBackend::Write(const Entry& entry) {
    SUB(Service, FS)                                                                               \
    SUB(Service, GRC)                                                                              \
    SUB(Service, HID)                                                                              \
+    SUB(Service, IRS)                                                                              \
    SUB(Service, LBL)                                                                              \
    SUB(Service, LDN)                                                                              \
    SUB(Service, LDR)                                                                              \
diff --git a/src/common/logging/log.h b/src/common/logging/log.h
index 4d577524f..abbd056ee 100644
--- a/src/common/logging/log.h
+++ b/src/common/logging/log.h
@@ -70,6 +70,7 @@ enum class Class : ClassType {
    Service_FS,        ///< The FS (Filesystem) service
    Service_GRC,       ///< The game recording service
    Service_HID,       ///< The HID (Human interface device) service
+    Service_IRS,       ///< The IRS service
    Service_LBL,       ///< The LBL (LCD backlight) service
    Service_LDN,       ///< The LDN (Local domain network) service
    Service_LDR,       ///< The loader service
diff --git a/src/common/ring_buffer.h b/src/common/ring_buffer.h
index 45926c9ec..abe3b4dc2 100644
--- a/src/common/ring_buffer.h
+++ b/src/common/ring_buffer.h
@@ -9,6 +9,7 @@
 #include <atomic>
 #include <cstddef>
 #include <cstring>
+#include <new>
 #include <type_traits>
 #include <vector>
 #include "common/common_types.h"
@@ -29,7 +30,7 @@ class RingBuffer {
    static_assert(capacity < std::numeric_limits<std::size_t>::max() / 2 / granularity);
    static_assert((capacity & (capacity - 1)) == 0, "capacity must be a power of two");
    // Ensure lock-free.
-    static_assert(std::atomic<std::size_t>::is_always_lock_free);
+    static_assert(std::atomic_size_t::is_always_lock_free);
 public:
    /// Pushes slots into the ring buffer
@@ -102,8 +103,15 @@ public:
 private:
    // It is important to align the below variables for performance reasons:
    // Having them on the same cache-line would result in false-sharing between them.
-    alignas(128) std::atomic<std::size_t> m_read_index{0};
+    // TODO: Remove this ifdef whenever clang and GCC support
-    alignas(128) std::atomic<std::size_t> m_write_index{0};
+    //       std::hardware_destructive_interference_size.
+#if defined(_MSC_VER) && _MSC_VER >= 1911
+    alignas(std::hardware_destructive_interference_size) std::atomic_size_t m_read_index{0};
+    alignas(std::hardware_destructive_interference_size) std::atomic_size_t m_write_index{0};
+#else
+    alignas(128) std::atomic_size_t m_read_index{0};
+    alignas(128) std::atomic_size_t m_write_index{0};
+#endif
    std::array<T, granularity * capacity> m_data;
 };
diff --git a/src/common/thread.h b/src/common/thread.h
index 12a1c095c..6cbdb96a3 100644
--- a/src/common/thread.h
+++ b/src/common/thread.h
@@ -87,14 +87,6 @@ private:
 void SleepCurrentThread(int ms);
 void SwitchCurrentThread(); // On Linux, this is equal to sleep 1ms
-// Use this function during a spin-wait to make the current thread
-// relax while another thread is working. This may be more efficient
-// than using events because event functions use kernel calls.
-inline void YieldCPU() {
-    std::this_thread::yield();
-}
 void SetCurrentThreadName(const char* name);
 } // namespace Common
diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt
index 26f727d96..23fd6e920 100644
--- a/src/core/CMakeLists.txt
+++ b/src/core/CMakeLists.txt
@@ -32,6 +32,8 @@ add_library(core STATIC
    file_sys/control_metadata.h
    file_sys/directory.h
    file_sys/errors.h
+    file_sys/fsmitm_romfsbuild.cpp
+    file_sys/fsmitm_romfsbuild.h
    file_sys/mode.h
    file_sys/nca_metadata.cpp
    file_sys/nca_metadata.h
@@ -59,10 +61,13 @@ add_library(core STATIC
    file_sys/vfs.h
    file_sys/vfs_concat.cpp
    file_sys/vfs_concat.h
+    file_sys/vfs_layered.cpp
+    file_sys/vfs_layered.h
    file_sys/vfs_offset.cpp
    file_sys/vfs_offset.h
    file_sys/vfs_real.cpp
    file_sys/vfs_real.h
+    file_sys/vfs_static.h
    file_sys/vfs_vector.cpp
    file_sys/vfs_vector.h
    file_sys/xts_archive.cpp
diff --git a/src/core/arm/arm_interface.h b/src/core/arm/arm_interface.h
index 867e34932..59da33f30 100644
--- a/src/core/arm/arm_interface.h
+++ b/src/core/arm/arm_interface.h
@@ -6,7 +6,10 @@
 #include <array>
 #include "common/common_types.h"
-#include "core/hle/kernel/vm_manager.h"
+namespace Kernel {
+enum class VMAPermission : u8;
+}
 namespace Core {
@@ -19,10 +22,16 @@ public:
        std::array<u64, 31> cpu_registers;
        u64 sp;
        u64 pc;
-        u64 pstate;
+        u32 pstate;
+        std::array<u8, 4> padding;
        std::array<u128, 32> vector_registers;
-        u64 fpcr;
+        u32 fpcr;
+        u32 fpsr;
+        u64 tpidr;
    };
+    // Internally within the kernel, it expects the AArch64 version of the
+    // thread context to be 800 bytes in size.
+    static_assert(sizeof(ThreadContext) == 0x320);
    /// Runs the CPU until an event happens
    virtual void Run() = 0;
diff --git a/src/core/arm/dynarmic/arm_dynarmic.cpp b/src/core/arm/dynarmic/arm_dynarmic.cpp
index 3f072c51f..05cc84458 100644
--- a/src/core/arm/dynarmic/arm_dynarmic.cpp
+++ b/src/core/arm/dynarmic/arm_dynarmic.cpp
@@ -12,8 +12,10 @@
 #include "core/core.h"
 #include "core/core_cpu.h"
 #include "core/core_timing.h"
+#include "core/gdbstub/gdbstub.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/svc.h"
+#include "core/hle/kernel/vm_manager.h"
 #include "core/memory.h"
 namespace Core {
@@ -79,6 +81,17 @@ public:
        case Dynarmic::A64::Exception::SendEventLocal:
        case Dynarmic::A64::Exception::Yield:
            return;
+        case Dynarmic::A64::Exception::Breakpoint:
+            if (GDBStub::IsServerEnabled()) {
+                parent.jit->HaltExecution();
+                parent.SetPC(pc);
+                Kernel::Thread* thread = Kernel::GetCurrentThread();
+                parent.SaveContext(thread->context);
+                GDBStub::Break();
+                GDBStub::SendTrap(thread, 5);
+                return;
+            }
+            [[fallthrough]];
        default:
            ASSERT_MSG(false, "ExceptionRaised(exception = {}, pc = {:X})",
                       static_cast<std::size_t>(exception), pc);
@@ -116,7 +129,8 @@ public:
 };
 std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const {
-    auto** const page_table = Core::CurrentProcess()->vm_manager.page_table.pointers.data();
+    auto& current_process = Core::CurrentProcess();
+    auto** const page_table = current_process->VMManager().page_table.pointers.data();
    Dynarmic::A64::UserConfig config;
@@ -125,7 +139,7 @@ std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const {
    // Memory
    config.page_table = reinterpret_cast<void**>(page_table);
-    config.page_table_address_space_bits = Memory::ADDRESS_SPACE_BITS;
+    config.page_table_address_space_bits = current_process->VMManager().GetAddressSpaceWidth();
    config.silently_mirror_page_table = false;
    // Multi-process state
@@ -161,7 +175,7 @@ ARM_Dynarmic::ARM_Dynarmic(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
                           std::size_t core_index)
    : cb(std::make_unique<ARM_Dynarmic_Callbacks>(*this)), core_index{core_index},
      exclusive_monitor{std::dynamic_pointer_cast<DynarmicExclusiveMonitor>(exclusive_monitor)} {
-    ThreadContext ctx;
+    ThreadContext ctx{};
    inner_unicorn.SaveContext(ctx);
    PageTableChanged();
    LoadContext(ctx);
@@ -233,15 +247,19 @@ void ARM_Dynarmic::SaveContext(ThreadContext& ctx) {
    ctx.pstate = jit->GetPstate();
    ctx.vector_registers = jit->GetVectors();
    ctx.fpcr = jit->GetFpcr();
+    ctx.fpsr = jit->GetFpsr();
+    ctx.tpidr = cb->tpidr_el0;
 }
 void ARM_Dynarmic::LoadContext(const ThreadContext& ctx) {
    jit->SetRegisters(ctx.cpu_registers);
    jit->SetSP(ctx.sp);
    jit->SetPC(ctx.pc);
-    jit->SetPstate(static_cast<u32>(ctx.pstate));
+    jit->SetPstate(ctx.pstate);
    jit->SetVectors(ctx.vector_registers);
-    jit->SetFpcr(static_cast<u32>(ctx.fpcr));
+    jit->SetFpcr(ctx.fpcr);
+    jit->SetFpsr(ctx.fpsr);
+    SetTPIDR_EL0(ctx.tpidr);
 }
 void ARM_Dynarmic::PrepareReschedule() {
diff --git a/src/core/arm/dynarmic/arm_dynarmic.h b/src/core/arm/dynarmic/arm_dynarmic.h
index e61382d3d..4ee92ee27 100644
--- a/src/core/arm/dynarmic/arm_dynarmic.h
+++ b/src/core/arm/dynarmic/arm_dynarmic.h
@@ -12,6 +12,10 @@
 #include "core/arm/exclusive_monitor.h"
 #include "core/arm/unicorn/arm_unicorn.h"
+namespace Memory {
+struct PageTable;
+}
 namespace Core {
 class ARM_Dynarmic_Callbacks;
diff --git a/src/core/core.cpp b/src/core/core.cpp
index 50f0a42fb..b6acfb3e4 100644
--- a/src/core/core.cpp
+++ b/src/core/core.cpp
@@ -64,7 +64,7 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
        if (concat.empty())
            return nullptr;
-        return FileSys::ConcatenateFiles(concat, dir->GetName());
+        return FileSys::ConcatenatedVfsFile::MakeConcatenatedFile(concat, dir->GetName());
    }
    return vfs->OpenFile(path, FileSys::Mode::Read);
@@ -202,7 +202,7 @@ struct System::Impl {
            return init_result;
        }
-        const Loader::ResultStatus load_result{app_loader->Load(kernel.CurrentProcess())};
+        const Loader::ResultStatus load_result{app_loader->Load(*kernel.CurrentProcess())};
        if (load_result != Loader::ResultStatus::Success) {
            LOG_CRITICAL(Core, "Failed to load ROM (Error {})!", static_cast<int>(load_result));
            Shutdown();
diff --git a/src/core/core_cpu.cpp b/src/core/core_cpu.cpp
index 21568ad50..265f8ed9c 100644
--- a/src/core/core_cpu.cpp
+++ b/src/core/core_cpu.cpp
@@ -55,16 +55,16 @@ Cpu::Cpu(std::shared_ptr<ExclusiveMonitor> exclusive_monitor,
    if (Settings::values.use_cpu_jit) {
 #ifdef ARCHITECTURE_x86_64
-        arm_interface = std::make_shared<ARM_Dynarmic>(exclusive_monitor, core_index);
+        arm_interface = std::make_unique<ARM_Dynarmic>(exclusive_monitor, core_index);
 #else
-        arm_interface = std::make_shared<ARM_Unicorn>();
+        arm_interface = std::make_unique<ARM_Unicorn>();
        LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
 #endif
    } else {
-        arm_interface = std::make_shared<ARM_Unicorn>();
+        arm_interface = std::make_unique<ARM_Unicorn>();
    }
-    scheduler = std::make_shared<Kernel::Scheduler>(arm_interface.get());
+    scheduler = std::make_shared<Kernel::Scheduler>(*arm_interface);
 }
 Cpu::~Cpu() = default;
diff --git a/src/core/core_cpu.h b/src/core/core_cpu.h
index 685532965..ee7e04abc 100644
--- a/src/core/core_cpu.h
+++ b/src/core/core_cpu.h
@@ -76,7 +76,7 @@ public:
 private:
    void Reschedule();
-    std::shared_ptr<ARM_Interface> arm_interface;
+    std::unique_ptr<ARM_Interface> arm_interface;
    std::shared_ptr<CpuBarrier> cpu_barrier;
    std::shared_ptr<Kernel::Scheduler> scheduler;
diff --git a/src/core/file_sys/bis_factory.cpp b/src/core/file_sys/bis_factory.cpp
index 205492897..6102ef476 100644
--- a/src/core/file_sys/bis_factory.cpp
+++ b/src/core/file_sys/bis_factory.cpp
@@ -2,13 +2,14 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+#include <fmt/format.h>
 #include "core/file_sys/bis_factory.h"
 #include "core/file_sys/registered_cache.h"
 namespace FileSys {
-BISFactory::BISFactory(VirtualDir nand_root_)
+BISFactory::BISFactory(VirtualDir nand_root_, VirtualDir load_root_)
-    : nand_root(std::move(nand_root_)),
+    : nand_root(std::move(nand_root_)), load_root(std::move(load_root_)),
      sysnand_cache(std::make_shared<RegisteredCache>(
          GetOrCreateDirectoryRelative(nand_root, "/system/Contents/registered"))),
      usrnand_cache(std::make_shared<RegisteredCache>(
@@ -24,4 +25,11 @@ std::shared_ptr<RegisteredCache> BISFactory::GetUserNANDContents() const {
    return usrnand_cache;
 }
+VirtualDir BISFactory::GetModificationLoadRoot(u64 title_id) const {
+    // LayeredFS doesn't work on updates and title id-less homebrew
+    if (title_id == 0 || (title_id & 0x800) > 0)
+        return nullptr;
+    return GetOrCreateDirectoryRelative(load_root, fmt::format("/{:016X}", title_id));
+}
 } // namespace FileSys
diff --git a/src/core/file_sys/bis_factory.h b/src/core/file_sys/bis_factory.h
index 9523dd864..c352e0925 100644
--- a/src/core/file_sys/bis_factory.h
+++ b/src/core/file_sys/bis_factory.h
@@ -17,14 +17,17 @@ class RegisteredCache;
 /// registered caches.
 class BISFactory {
 public:
-    explicit BISFactory(VirtualDir nand_root);
+    explicit BISFactory(VirtualDir nand_root, VirtualDir load_root);
    ~BISFactory();
    std::shared_ptr<RegisteredCache> GetSystemNANDContents() const;
    std::shared_ptr<RegisteredCache> GetUserNANDContents() const;
+    VirtualDir GetModificationLoadRoot(u64 title_id) const;
 private:
    VirtualDir nand_root;
+    VirtualDir load_root;
    std::shared_ptr<RegisteredCache> sysnand_cache;
    std::shared_ptr<RegisteredCache> usrnand_cache;
diff --git a/src/core/file_sys/content_archive.cpp b/src/core/file_sys/content_archive.cpp
index 45fc0b574..aa1b3c17d 100644
--- a/src/core/file_sys/content_archive.cpp
+++ b/src/core/file_sys/content_archive.cpp
@@ -463,6 +463,8 @@ NCA::NCA(VirtualFile file_, VirtualFile bktr_base_romfs_, u64 bktr_base_ivfc_off
    status = Loader::ResultStatus::Success;
 }
+NCA::~NCA() = default;
 Loader::ResultStatus NCA::GetStatus() const {
    return status;
 }
diff --git a/src/core/file_sys/content_archive.h b/src/core/file_sys/content_archive.h
index 00eca52da..f9f66cae9 100644
--- a/src/core/file_sys/content_archive.h
+++ b/src/core/file_sys/content_archive.h
@@ -81,6 +81,8 @@ class NCA : public ReadOnlyVfsDirectory {
 public:
    explicit NCA(VirtualFile file, VirtualFile bktr_base_romfs = nullptr,
                 u64 bktr_base_ivfc_offset = 0);
+    ~NCA() override;
    Loader::ResultStatus GetStatus() const;
    std::vector<std::shared_ptr<VfsFile>> GetFiles() const override;
diff --git a/src/core/file_sys/control_metadata.cpp b/src/core/file_sys/control_metadata.cpp
index e76bf77bf..5b1177a03 100644
--- a/src/core/file_sys/control_metadata.cpp
+++ b/src/core/file_sys/control_metadata.cpp
@@ -8,6 +8,14 @@
 namespace FileSys {
+const std::array<const char*, 15> LANGUAGE_NAMES = {
+    "AmericanEnglish", "BritishEnglish", "Japanese",
+    "French",          "German",         "LatinAmericanSpanish",
+    "Spanish",         "Italian",        "Dutch",
+    "CanadianFrench",  "Portugese",      "Russian",
+    "Korean",          "Taiwanese",      "Chinese",
+};
 std::string LanguageEntry::GetApplicationName() const {
    return Common::StringFromFixedZeroTerminatedBuffer(application_name.data(), 0x200);
 }
@@ -20,18 +28,20 @@ NACP::NACP(VirtualFile file) : raw(std::make_unique<RawNACP>()) {
    file->ReadObject(raw.get());
 }
+NACP::~NACP() = default;
 const LanguageEntry& NACP::GetLanguageEntry(Language language) const {
    if (language != Language::Default) {
        return raw->language_entries.at(static_cast<u8>(language));
-    } else {
-        for (const auto& language_entry : raw->language_entries) {
-            if (!language_entry.GetApplicationName().empty())
-                return language_entry;
-        }
-        // Fallback to English
-        return GetLanguageEntry(Language::AmericanEnglish);
    }
+    for (const auto& language_entry : raw->language_entries) {
+        if (!language_entry.GetApplicationName().empty())
+            return language_entry;
+    }
+    // Fallback to English
+    return GetLanguageEntry(Language::AmericanEnglish);
 }
 std::string NACP::GetApplicationName(Language language) const {
diff --git a/src/core/file_sys/control_metadata.h b/src/core/file_sys/control_metadata.h
index 8a510bf46..43d6f0719 100644
--- a/src/core/file_sys/control_metadata.h
+++ b/src/core/file_sys/control_metadata.h
@@ -66,18 +66,15 @@ enum class Language : u8 {
    Default = 255,
 };
-static constexpr std::array<const char*, 15> LANGUAGE_NAMES = {
+extern const std::array<const char*, 15> LANGUAGE_NAMES;
-    "AmericanEnglish", "BritishEnglish", "Japanese",
-    "French",          "German",         "LatinAmericanSpanish",
-    "Spanish",         "Italian",        "Dutch",
-    "CanadianFrench",  "Portugese",      "Russian",
-    "Korean",          "Taiwanese",      "Chinese"};
 // A class representing the format used by NX metadata files, typically named Control.nacp.
 // These store application name, dev name, title id, and other miscellaneous data.
 class NACP {
 public:
    explicit NACP(VirtualFile file);
+    ~NACP();
    const LanguageEntry& GetLanguageEntry(Language language = Language::Default) const;
    std::string GetApplicationName(Language language = Language::Default) const;
    std::string GetDeveloperName(Language language = Language::Default) const;
diff --git a/src/core/file_sys/fsmitm_romfsbuild.cpp b/src/core/file_sys/fsmitm_romfsbuild.cpp
new file mode 100644
index 000000000..2a913ce82
--- /dev/null
+++ b/src/core/file_sys/fsmitm_romfsbuild.cpp
@@ -0,0 +1,366 @@
+/*
+ * Copyright (c) 2018 Atmosphère-NX
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+/*
+ * Adapted by DarkLordZach for use/interaction with yuzu
+ *
+ * Modifications Copyright 2018 yuzu emulator team
+ * Licensed under GPLv2 or any later version
+ * Refer to the license.txt file included.
+ */
+#include <cstring>
+#include "common/alignment.h"
+#include "common/assert.h"
+#include "core/file_sys/fsmitm_romfsbuild.h"
+#include "core/file_sys/vfs.h"
+#include "core/file_sys/vfs_vector.h"
+namespace FileSys {
+constexpr u64 FS_MAX_PATH = 0x301;
+constexpr u32 ROMFS_ENTRY_EMPTY = 0xFFFFFFFF;
+constexpr u32 ROMFS_FILEPARTITION_OFS = 0x200;
+// Types for building a RomFS.
+struct RomFSHeader {
+    u64 header_size;
+    u64 dir_hash_table_ofs;
+    u64 dir_hash_table_size;
+    u64 dir_table_ofs;
+    u64 dir_table_size;
+    u64 file_hash_table_ofs;
+    u64 file_hash_table_size;
+    u64 file_table_ofs;
+    u64 file_table_size;
+    u64 file_partition_ofs;
+};
+static_assert(sizeof(RomFSHeader) == 0x50, "RomFSHeader has incorrect size.");
+struct RomFSDirectoryEntry {
+    u32 parent;
+    u32 sibling;
+    u32 child;
+    u32 file;
+    u32 hash;
+    u32 name_size;
+};
+static_assert(sizeof(RomFSDirectoryEntry) == 0x18, "RomFSDirectoryEntry has incorrect size.");
+struct RomFSFileEntry {
+    u32 parent;
+    u32 sibling;
+    u64 offset;
+    u64 size;
+    u32 hash;
+    u32 name_size;
+};
+static_assert(sizeof(RomFSFileEntry) == 0x20, "RomFSFileEntry has incorrect size.");
+struct RomFSBuildFileContext;
+struct RomFSBuildDirectoryContext {
+    std::string path;
+    u32 cur_path_ofs = 0;
+    u32 path_len = 0;
+    u32 entry_offset = 0;
+    std::shared_ptr<RomFSBuildDirectoryContext> parent;
+    std::shared_ptr<RomFSBuildDirectoryContext> child;
+    std::shared_ptr<RomFSBuildDirectoryContext> sibling;
+    std::shared_ptr<RomFSBuildFileContext> file;
+};
+struct RomFSBuildFileContext {
+    std::string path;
+    u32 cur_path_ofs = 0;
+    u32 path_len = 0;
+    u32 entry_offset = 0;
+    u64 offset = 0;
+    u64 size = 0;
+    std::shared_ptr<RomFSBuildDirectoryContext> parent;
+    std::shared_ptr<RomFSBuildFileContext> sibling;
+    VirtualFile source;
+};
+static u32 romfs_calc_path_hash(u32 parent, std::string path, u32 start, std::size_t path_len) {
+    u32 hash = parent ^ 123456789;
+    for (u32 i = 0; i < path_len; i++) {
+        hash = (hash >> 5) | (hash << 27);
+        hash ^= path[start + i];
+    }
+    return hash;
+}
+static u64 romfs_get_hash_table_count(u64 num_entries) {
+    if (num_entries < 3) {
+        return 3;
+    }
+    if (num_entries < 19) {
+        return num_entries | 1;
+    }
+    u64 count = num_entries;
+    while (count % 2 == 0 || count % 3 == 0 || count % 5 == 0 || count % 7 == 0 ||
+           count % 11 == 0 || count % 13 == 0 || count % 17 == 0) {
+        count++;
+    }
+    return count;
+}
+void RomFSBuildContext::VisitDirectory(VirtualDir root_romfs,
+                                       std::shared_ptr<RomFSBuildDirectoryContext> parent) {
+    std::vector<std::shared_ptr<RomFSBuildDirectoryContext>> child_dirs;
+    VirtualDir dir;
+    if (parent->path_len == 0)
+        dir = root_romfs;
+    else
+        dir = root_romfs->GetDirectoryRelative(parent->path);
+    const auto entries = dir->GetEntries();
+    for (const auto& kv : entries) {
+        if (kv.second == VfsEntryType::Directory) {
+            const auto child = std::make_shared<RomFSBuildDirectoryContext>();
+            // Set child's path.
+            child->cur_path_ofs = parent->path_len + 1;
+            child->path_len = child->cur_path_ofs + static_cast<u32>(kv.first.size());
+            child->path = parent->path + "/" + kv.first;
+            // Sanity check on path_len
+            ASSERT(child->path_len < FS_MAX_PATH);
+            if (AddDirectory(parent, child)) {
+                child_dirs.push_back(child);
+            }
+        } else {
+            const auto child = std::make_shared<RomFSBuildFileContext>();
+            // Set child's path.
+            child->cur_path_ofs = parent->path_len + 1;
+            child->path_len = child->cur_path_ofs + static_cast<u32>(kv.first.size());
+            child->path = parent->path + "/" + kv.first;
+            // Sanity check on path_len
+            ASSERT(child->path_len < FS_MAX_PATH);
+            child->source = root_romfs->GetFileRelative(child->path);
+            child->size = child->source->GetSize();
+            AddFile(parent, child);
+        }
+    }
+    for (auto& child : child_dirs) {
+        this->VisitDirectory(root_romfs, child);
+    }
+}
+bool RomFSBuildContext::AddDirectory(std::shared_ptr<RomFSBuildDirectoryContext> parent_dir_ctx,
+                                     std::shared_ptr<RomFSBuildDirectoryContext> dir_ctx) {
+    // Check whether it's already in the known directories.
+    const auto existing = directories.find(dir_ctx->path);
+    if (existing != directories.end())
+        return false;
+    // Add a new directory.
+    num_dirs++;
+    dir_table_size +=
+        sizeof(RomFSDirectoryEntry) + Common::AlignUp(dir_ctx->path_len - dir_ctx->cur_path_ofs, 4);
+    dir_ctx->parent = parent_dir_ctx;
+    directories.emplace(dir_ctx->path, dir_ctx);
+    return true;
+}
+bool RomFSBuildContext::AddFile(std::shared_ptr<RomFSBuildDirectoryContext> parent_dir_ctx,
+                                std::shared_ptr<RomFSBuildFileContext> file_ctx) {
+    // Check whether it's already in the known files.
+    const auto existing = files.find(file_ctx->path);
+    if (existing != files.end()) {
+        return false;
+    }
+    // Add a new file.
+    num_files++;
+    file_table_size +=
+        sizeof(RomFSFileEntry) + Common::AlignUp(file_ctx->path_len - file_ctx->cur_path_ofs, 4);
+    file_ctx->parent = parent_dir_ctx;
+    files.emplace(file_ctx->path, file_ctx);
+    return true;
+}
+RomFSBuildContext::RomFSBuildContext(VirtualDir base_) : base(std::move(base_)) {
+    root = std::make_shared<RomFSBuildDirectoryContext>();
+    root->path = "\0";
+    directories.emplace(root->path, root);
+    num_dirs = 1;
+    dir_table_size = 0x18;
+    VisitDirectory(base, root);
+}
+RomFSBuildContext::~RomFSBuildContext() = default;
+std::map<u64, VirtualFile> RomFSBuildContext::Build() {
+    const u64 dir_hash_table_entry_count = romfs_get_hash_table_count(num_dirs);
+    const u64 file_hash_table_entry_count = romfs_get_hash_table_count(num_files);
+    dir_hash_table_size = 4 * dir_hash_table_entry_count;
+    file_hash_table_size = 4 * file_hash_table_entry_count;
+    // Assign metadata pointers
+    RomFSHeader header{};
+    std::vector<u32> dir_hash_table(dir_hash_table_entry_count, ROMFS_ENTRY_EMPTY);
+    std::vector<u32> file_hash_table(file_hash_table_entry_count, ROMFS_ENTRY_EMPTY);
+    std::vector<u8> dir_table(dir_table_size);
+    std::vector<u8> file_table(file_table_size);
+    std::shared_ptr<RomFSBuildFileContext> cur_file;
+    // Determine file offsets.
+    u32 entry_offset = 0;
+    std::shared_ptr<RomFSBuildFileContext> prev_file = nullptr;
+    for (const auto& it : files) {
+        cur_file = it.second;
+        file_partition_size = Common::AlignUp(file_partition_size, 16);
+        cur_file->offset = file_partition_size;
+        file_partition_size += cur_file->size;
+        cur_file->entry_offset = entry_offset;
+        entry_offset += sizeof(RomFSFileEntry) +
+                        Common::AlignUp(cur_file->path_len - cur_file->cur_path_ofs, 4);
+        prev_file = cur_file;
+    }
+    // Assign deferred parent/sibling ownership.
+    for (auto it = files.rbegin(); it != files.rend(); ++it) {
+        cur_file = it->second;
+        cur_file->sibling = cur_file->parent->file;
+        cur_file->parent->file = cur_file;
+    }
+    std::shared_ptr<RomFSBuildDirectoryContext> cur_dir;
+    // Determine directory offsets.
+    entry_offset = 0;
+    for (const auto& it : directories) {
+        cur_dir = it.second;
+        cur_dir->entry_offset = entry_offset;
+        entry_offset += sizeof(RomFSDirectoryEntry) +
+                        Common::AlignUp(cur_dir->path_len - cur_dir->cur_path_ofs, 4);
+    }
+    // Assign deferred parent/sibling ownership.
+    for (auto it = directories.rbegin(); it->second != root; ++it) {
+        cur_dir = it->second;
+        cur_dir->sibling = cur_dir->parent->child;
+        cur_dir->parent->child = cur_dir;
+    }
+    std::map<u64, VirtualFile> out;
+    // Populate file tables.
+    for (const auto& it : files) {
+        cur_file = it.second;
+        RomFSFileEntry cur_entry{};
+        cur_entry.parent = cur_file->parent->entry_offset;
+        cur_entry.sibling =
+            cur_file->sibling == nullptr ? ROMFS_ENTRY_EMPTY : cur_file->sibling->entry_offset;
+        cur_entry.offset = cur_file->offset;
+        cur_entry.size = cur_file->size;
+        const auto name_size = cur_file->path_len - cur_file->cur_path_ofs;
+        const auto hash = romfs_calc_path_hash(cur_file->parent->entry_offset, cur_file->path,
+                                               cur_file->cur_path_ofs, name_size);
+        cur_entry.hash = file_hash_table[hash % file_hash_table_entry_count];
+        file_hash_table[hash % file_hash_table_entry_count] = cur_file->entry_offset;
+        cur_entry.name_size = name_size;
+        out.emplace(cur_file->offset + ROMFS_FILEPARTITION_OFS, cur_file->source);
+        std::memcpy(file_table.data() + cur_file->entry_offset, &cur_entry, sizeof(RomFSFileEntry));
+        std::memset(file_table.data() + cur_file->entry_offset + sizeof(RomFSFileEntry), 0,
+                    Common::AlignUp(cur_entry.name_size, 4));
+        std::memcpy(file_table.data() + cur_file->entry_offset + sizeof(RomFSFileEntry),
+                    cur_file->path.data() + cur_file->cur_path_ofs, name_size);
+    }
+    // Populate dir tables.
+    for (const auto& it : directories) {
+        cur_dir = it.second;
+        RomFSDirectoryEntry cur_entry{};
+        cur_entry.parent = cur_dir == root ? 0 : cur_dir->parent->entry_offset;
+        cur_entry.sibling =
+            cur_dir->sibling == nullptr ? ROMFS_ENTRY_EMPTY : cur_dir->sibling->entry_offset;
+        cur_entry.child =
+            cur_dir->child == nullptr ? ROMFS_ENTRY_EMPTY : cur_dir->child->entry_offset;
+        cur_entry.file = cur_dir->file == nullptr ? ROMFS_ENTRY_EMPTY : cur_dir->file->entry_offset;
+        const auto name_size = cur_dir->path_len - cur_dir->cur_path_ofs;
+        const auto hash = romfs_calc_path_hash(cur_dir == root ? 0 : cur_dir->parent->entry_offset,
+                                               cur_dir->path, cur_dir->cur_path_ofs, name_size);
+        cur_entry.hash = dir_hash_table[hash % dir_hash_table_entry_count];
+        dir_hash_table[hash % dir_hash_table_entry_count] = cur_dir->entry_offset;
+        cur_entry.name_size = name_size;
+        std::memcpy(dir_table.data() + cur_dir->entry_offset, &cur_entry,
+                    sizeof(RomFSDirectoryEntry));
+        std::memset(dir_table.data() + cur_dir->entry_offset + sizeof(RomFSDirectoryEntry), 0,
+                    Common::AlignUp(cur_entry.name_size, 4));
+        std::memcpy(dir_table.data() + cur_dir->entry_offset + sizeof(RomFSDirectoryEntry),
+                    cur_dir->path.data() + cur_dir->cur_path_ofs, name_size);
+    }
+    // Set header fields.
+    header.header_size = sizeof(RomFSHeader);
+    header.file_hash_table_size = file_hash_table_size;
+    header.file_table_size = file_table_size;
+    header.dir_hash_table_size = dir_hash_table_size;
+    header.dir_table_size = dir_table_size;
+    header.file_partition_ofs = ROMFS_FILEPARTITION_OFS;
+    header.dir_hash_table_ofs = Common::AlignUp(header.file_partition_ofs + file_partition_size, 4);
+    header.dir_table_ofs = header.dir_hash_table_ofs + header.dir_hash_table_size;
+    header.file_hash_table_ofs = header.dir_table_ofs + header.dir_table_size;
+    header.file_table_ofs = header.file_hash_table_ofs + header.file_hash_table_size;
+    std::vector<u8> header_data(sizeof(RomFSHeader));
+    std::memcpy(header_data.data(), &header, header_data.size());
+    out.emplace(0, std::make_shared<VectorVfsFile>(std::move(header_data)));
+    std::vector<u8> metadata(file_hash_table_size + file_table_size + dir_hash_table_size +
+                             dir_table_size);
+    std::size_t index = 0;
+    std::memcpy(metadata.data(), dir_hash_table.data(), dir_hash_table.size() * sizeof(u32));
+    index += dir_hash_table.size() * sizeof(u32);
+    std::memcpy(metadata.data() + index, dir_table.data(), dir_table.size());
+    index += dir_table.size();
+    std::memcpy(metadata.data() + index, file_hash_table.data(),
+                file_hash_table.size() * sizeof(u32));
+    index += file_hash_table.size() * sizeof(u32);
+    std::memcpy(metadata.data() + index, file_table.data(), file_table.size());
+    out.emplace(header.dir_hash_table_ofs, std::make_shared<VectorVfsFile>(std::move(metadata)));
+    return out;
+}
+} // namespace FileSys
diff --git a/src/core/file_sys/fsmitm_romfsbuild.h b/src/core/file_sys/fsmitm_romfsbuild.h
new file mode 100644
index 000000000..b0c3c123b
--- /dev/null
+++ b/src/core/file_sys/fsmitm_romfsbuild.h
@@ -0,0 +1,70 @@
+/*
+ * Copyright (c) 2018 Atmosphère-NX
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+/*
+ * Adapted by DarkLordZach for use/interaction with yuzu
+ *
+ * Modifications Copyright 2018 yuzu emulator team
+ * Licensed under GPLv2 or any later version
+ * Refer to the license.txt file included.
+ */
+#pragma once
+#include <map>
+#include <memory>
+#include <string>
+#include <boost/detail/container_fwd.hpp>
+#include "common/common_types.h"
+#include "core/file_sys/vfs.h"
+namespace FileSys {
+struct RomFSBuildDirectoryContext;
+struct RomFSBuildFileContext;
+struct RomFSDirectoryEntry;
+struct RomFSFileEntry;
+class RomFSBuildContext {
+public:
+    explicit RomFSBuildContext(VirtualDir base);
+    ~RomFSBuildContext();
+    // This finalizes the context.
+    std::map<u64, VirtualFile> Build();
+private:
+    VirtualDir base;
+    std::shared_ptr<RomFSBuildDirectoryContext> root;
+    std::map<std::string, std::shared_ptr<RomFSBuildDirectoryContext>, std::less<>> directories;
+    std::map<std::string, std::shared_ptr<RomFSBuildFileContext>, std::less<>> files;
+    u64 num_dirs = 0;
+    u64 num_files = 0;
+    u64 dir_table_size = 0;
+    u64 file_table_size = 0;
+    u64 dir_hash_table_size = 0;
+    u64 file_hash_table_size = 0;
+    u64 file_partition_size = 0;
+    void VisitDirectory(VirtualDir filesys, std::shared_ptr<RomFSBuildDirectoryContext> parent);
+    bool AddDirectory(std::shared_ptr<RomFSBuildDirectoryContext> parent_dir_ctx,
+                      std::shared_ptr<RomFSBuildDirectoryContext> dir_ctx);
+    bool AddFile(std::shared_ptr<RomFSBuildDirectoryContext> parent_dir_ctx,
+                 std::shared_ptr<RomFSBuildFileContext> file_ctx);
+};
+} // namespace FileSys
diff --git a/src/core/file_sys/nca_metadata.cpp b/src/core/file_sys/nca_metadata.cpp
index 479916b69..6f34b7836 100644
--- a/src/core/file_sys/nca_metadata.cpp
+++ b/src/core/file_sys/nca_metadata.cpp
@@ -51,6 +51,8 @@ CNMT::CNMT(CNMTHeader header, OptionalHeader opt_header, std::vector<ContentReco
    : header(std::move(header)), opt_header(std::move(opt_header)),
      content_records(std::move(content_records)), meta_records(std::move(meta_records)) {}
+CNMT::~CNMT() = default;
 u64 CNMT::GetTitleID() const {
    return header.title_id;
 }
diff --git a/src/core/file_sys/nca_metadata.h b/src/core/file_sys/nca_metadata.h
index da5a8dbe8..a05d155f4 100644
--- a/src/core/file_sys/nca_metadata.h
+++ b/src/core/file_sys/nca_metadata.h
@@ -87,6 +87,7 @@ public:
    explicit CNMT(VirtualFile file);
    CNMT(CNMTHeader header, OptionalHeader opt_header, std::vector<ContentRecord> content_records,
         std::vector<MetaRecord> meta_records);
+    ~CNMT();
    u64 GetTitleID() const;
    u32 GetTitleVersion() const;
diff --git a/src/core/file_sys/partition_filesystem.cpp b/src/core/file_sys/partition_filesystem.cpp
index f5b3b0175..5791c76ff 100644
--- a/src/core/file_sys/partition_filesystem.cpp
+++ b/src/core/file_sys/partition_filesystem.cpp
@@ -72,6 +72,8 @@ PartitionFilesystem::PartitionFilesystem(std::shared_ptr<VfsFile> file) {
    status = Loader::ResultStatus::Success;
 }
+PartitionFilesystem::~PartitionFilesystem() = default;
 Loader::ResultStatus PartitionFilesystem::GetStatus() const {
    return status;
 }
diff --git a/src/core/file_sys/partition_filesystem.h b/src/core/file_sys/partition_filesystem.h
index e80d2456b..739c63a7f 100644
--- a/src/core/file_sys/partition_filesystem.h
+++ b/src/core/file_sys/partition_filesystem.h
@@ -25,6 +25,8 @@ namespace FileSys {
 class PartitionFilesystem : public ReadOnlyVfsDirectory {
 public:
    explicit PartitionFilesystem(std::shared_ptr<VfsFile> file);
+    ~PartitionFilesystem() override;
    Loader::ResultStatus GetStatus() const;
    std::vector<std::shared_ptr<VfsFile>> GetFiles() const override;
diff --git a/src/core/file_sys/patch_manager.cpp b/src/core/file_sys/patch_manager.cpp
index b37b4c68b..4b3b5e665 100644
--- a/src/core/file_sys/patch_manager.cpp
+++ b/src/core/file_sys/patch_manager.cpp
@@ -11,6 +11,7 @@
 #include "core/file_sys/patch_manager.h"
 #include "core/file_sys/registered_cache.h"
 #include "core/file_sys/romfs.h"
+#include "core/file_sys/vfs_layered.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/loader/loader.h"
@@ -31,8 +32,9 @@ std::string FormatTitleVersion(u32 version, TitleVersionFormat format) {
    return fmt::format("v{}.{}.{}", bytes[3], bytes[2], bytes[1]);
 }
-constexpr std::array<const char*, 1> PATCH_TYPE_NAMES{
+constexpr std::array<const char*, 2> PATCH_TYPE_NAMES{
    "Update",
+    "LayeredFS",
 };
 std::string FormatPatchTypeName(PatchType type) {
@@ -41,6 +43,8 @@ std::string FormatPatchTypeName(PatchType type) {
 PatchManager::PatchManager(u64 title_id) : title_id(title_id) {}
+PatchManager::~PatchManager() = default;
 VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
    LOG_INFO(Loader, "Patching ExeFS for title_id={:016X}", title_id);
@@ -64,6 +68,44 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
    return exefs;
 }
+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) {
+        return;
+    }
+    auto extracted = ExtractRomFS(romfs);
+    if (extracted == nullptr) {
+        return;
+    }
+    auto patch_dirs = load_dir->GetSubdirectories();
+    std::sort(patch_dirs.begin(), patch_dirs.end(),
+              [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });
+    std::vector<VirtualDir> layers;
+    layers.reserve(patch_dirs.size() + 1);
+    for (const auto& subdir : patch_dirs) {
+        auto romfs_dir = subdir->GetSubdirectory("romfs");
+        if (romfs_dir != nullptr)
+            layers.push_back(std::move(romfs_dir));
+    }
+    layers.push_back(std::move(extracted));
+    auto layered = LayeredVfsDirectory::MakeLayeredDirectory(std::move(layers));
+    if (layered == nullptr) {
+        return;
+    }
+    auto packed = CreateRomFS(std::move(layered));
+    if (packed == nullptr) {
+        return;
+    }
+    LOG_INFO(Loader, "    RomFS: LayeredFS patches applied successfully");
+    romfs = std::move(packed);
+}
 VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset,
                                     ContentRecordType type) const {
    LOG_INFO(Loader, "Patching RomFS for title_id={:016X}, type={:02X}", title_id,
@@ -87,6 +129,9 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset,
        }
    }
+    // LayeredFS
+    ApplyLayeredFS(romfs, title_id, type);
    return romfs;
 }
@@ -112,6 +157,10 @@ std::map<PatchType, std::string> PatchManager::GetPatchVersionNames() const {
        }
    }
+    const auto lfs_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
+    if (lfs_dir != nullptr && lfs_dir->GetSize() > 0)
+        out.insert_or_assign(PatchType::LayeredFS, "");
    return out;
 }
diff --git a/src/core/file_sys/patch_manager.h b/src/core/file_sys/patch_manager.h
index b521977b2..464f17515 100644
--- a/src/core/file_sys/patch_manager.h
+++ b/src/core/file_sys/patch_manager.h
@@ -26,6 +26,7 @@ std::string FormatTitleVersion(u32 version,
 enum class PatchType {
    Update,
+    LayeredFS,
 };
 std::string FormatPatchTypeName(PatchType type);
@@ -34,6 +35,7 @@ std::string FormatPatchTypeName(PatchType type);
 class PatchManager {
 public:
    explicit PatchManager(u64 title_id);
+    ~PatchManager();
    // Currently tracked ExeFS patches:
    // - Game Updates
@@ -41,6 +43,7 @@ public:
    // Currently tracked RomFS patches:
    // - Game Updates
+    // - LayeredFS
    VirtualFile PatchRomFS(VirtualFile base, u64 ivfc_offset,
                           ContentRecordType type = ContentRecordType::Program) const;
diff --git a/src/core/file_sys/program_metadata.cpp b/src/core/file_sys/program_metadata.cpp
index 9d19aaa6d..8903ed1d3 100644
--- a/src/core/file_sys/program_metadata.cpp
+++ b/src/core/file_sys/program_metadata.cpp
@@ -12,6 +12,10 @@
 namespace FileSys {
+ProgramMetadata::ProgramMetadata() = default;
+ProgramMetadata::~ProgramMetadata() = default;
 Loader::ResultStatus ProgramMetadata::Load(VirtualFile file) {
    std::size_t total_size = static_cast<std::size_t>(file->GetSize());
    if (total_size < sizeof(Header))
@@ -79,10 +83,12 @@ void ProgramMetadata::Print() const {
    auto address_space = "Unknown";
    switch (npdm_header.address_space_type) {
-    case ProgramAddressSpaceType::Is64Bit:
+    case ProgramAddressSpaceType::Is36Bit:
+    case ProgramAddressSpaceType::Is39Bit:
        address_space = "64-bit";
        break;
    case ProgramAddressSpaceType::Is32Bit:
+    case ProgramAddressSpaceType::Is32BitNoMap:
        address_space = "32-bit";
        break;
    }
diff --git a/src/core/file_sys/program_metadata.h b/src/core/file_sys/program_metadata.h
index 3c0a49f16..e4470d6f0 100644
--- a/src/core/file_sys/program_metadata.h
+++ b/src/core/file_sys/program_metadata.h
@@ -17,8 +17,10 @@ enum class ResultStatus : u16;
 namespace FileSys {
 enum class ProgramAddressSpaceType : u8 {
-    Is64Bit = 1,
+    Is32Bit = 0,
-    Is32Bit = 2,
+    Is36Bit = 1,
+    Is32BitNoMap = 2,
+    Is39Bit = 3,
 };
 enum class ProgramFilePermission : u64 {
@@ -36,6 +38,9 @@ enum class ProgramFilePermission : u64 {
 */
 class ProgramMetadata {
 public:
+    ProgramMetadata();
+    ~ProgramMetadata();
    Loader::ResultStatus Load(VirtualFile file);
    bool Is64BitProgram() const;
diff --git a/src/core/file_sys/registered_cache.cpp b/src/core/file_sys/registered_cache.cpp
index dad7ae10b..e9b040689 100644
--- a/src/core/file_sys/registered_cache.cpp
+++ b/src/core/file_sys/registered_cache.cpp
@@ -18,6 +18,10 @@
 #include "core/loader/loader.h"
 namespace FileSys {
+// The size of blocks to use when vfs raw copying into nand.
+constexpr size_t VFS_RC_LARGE_COPY_BLOCK = 0x400000;
 std::string RegisteredCacheEntry::DebugInfo() const {
    return fmt::format("title_id={:016X}, content_type={:02X}", title_id, static_cast<u8>(type));
 }
@@ -121,7 +125,7 @@ VirtualFile RegisteredCache::OpenFileOrDirectoryConcat(const VirtualDir& dir,
            if (concat.empty())
                return nullptr;
-            file = FileSys::ConcatenateFiles(concat);
+            file = ConcatenatedVfsFile::MakeConcatenatedFile(concat, concat.front()->GetName());
        }
        return file;
@@ -480,7 +484,8 @@ InstallResult RegisteredCache::RawInstallNCA(std::shared_ptr<NCA> nca, const Vfs
    auto out = dir->CreateFileRelative(path);
    if (out == nullptr)
        return InstallResult::ErrorCopyFailed;
-    return copy(in, out) ? InstallResult::Success : InstallResult::ErrorCopyFailed;
+    return copy(in, out, VFS_RC_LARGE_COPY_BLOCK) ? InstallResult::Success
+                                                  : InstallResult::ErrorCopyFailed;
 }
 bool RegisteredCache::RawInstallYuzuMeta(const CNMT& cnmt) {
diff --git a/src/core/file_sys/registered_cache.h b/src/core/file_sys/registered_cache.h
index f487b0cf0..c0cd59fc5 100644
--- a/src/core/file_sys/registered_cache.h
+++ b/src/core/file_sys/registered_cache.h
@@ -27,7 +27,7 @@ struct ContentRecord;
 using NcaID = std::array<u8, 0x10>;
 using RegisteredCacheParsingFunction = std::function<VirtualFile(const VirtualFile&, const NcaID&)>;
-using VfsCopyFunction = std::function<bool(VirtualFile, VirtualFile)>;
+using VfsCopyFunction = std::function<bool(const VirtualFile&, const VirtualFile&, size_t)>;
 enum class InstallResult {
    Success,
diff --git a/src/core/file_sys/romfs.cpp b/src/core/file_sys/romfs.cpp
index 9f6e41cdf..5910f7046 100644
--- a/src/core/file_sys/romfs.cpp
+++ b/src/core/file_sys/romfs.cpp
@@ -4,8 +4,10 @@
 #include "common/common_types.h"
 #include "common/swap.h"
+#include "core/file_sys/fsmitm_romfsbuild.h"
 #include "core/file_sys/romfs.h"
 #include "core/file_sys/vfs.h"
+#include "core/file_sys/vfs_concat.h"
 #include "core/file_sys/vfs_offset.h"
 #include "core/file_sys/vfs_vector.h"
@@ -98,7 +100,7 @@ void ProcessDirectory(VirtualFile file, std::size_t dir_offset, std::size_t file
    }
 }
-VirtualDir ExtractRomFS(VirtualFile file) {
+VirtualDir ExtractRomFS(VirtualFile file, RomFSExtractionType type) {
    RomFSHeader header{};
    if (file->ReadObject(&header) != sizeof(RomFSHeader))
        return nullptr;
@@ -117,9 +119,22 @@ VirtualDir ExtractRomFS(VirtualFile file) {
    VirtualDir out = std::move(root);
-    while (out->GetSubdirectory("") != nullptr)
+    while (out->GetSubdirectories().size() == 1 && out->GetFiles().empty()) {
-        out = out->GetSubdirectory("");
+        if (out->GetSubdirectories().front()->GetName() == "data" &&
+            type == RomFSExtractionType::Truncated)
+            break;
+        out = out->GetSubdirectories().front();
+    }
    return out;
 }
+VirtualFile CreateRomFS(VirtualDir dir) {
+    if (dir == nullptr)
+        return nullptr;
+    RomFSBuildContext ctx{dir};
+    return ConcatenatedVfsFile::MakeConcatenatedFile(0, ctx.Build(), dir->GetName());
+}
 } // namespace FileSys
diff --git a/src/core/file_sys/romfs.h b/src/core/file_sys/romfs.h
index e54a7d7a9..ecd1eb725 100644
--- a/src/core/file_sys/romfs.h
+++ b/src/core/file_sys/romfs.h
@@ -5,6 +5,7 @@
 #pragma once
 #include <array>
+#include <map>
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/swap.h"
@@ -12,6 +13,8 @@
 namespace FileSys {
+struct RomFSHeader;
 struct IVFCLevel {
    u64_le offset;
    u64_le size;
@@ -29,8 +32,18 @@ struct IVFCHeader {
 };
 static_assert(sizeof(IVFCHeader) == 0xE0, "IVFCHeader has incorrect size.");
+enum class RomFSExtractionType {
+    Full,      // Includes data directory
+    Truncated, // Traverses into data directory
+};
 // Converts a RomFS binary blob to VFS Filesystem
 // Returns nullptr on failure
-VirtualDir ExtractRomFS(VirtualFile file);
+VirtualDir ExtractRomFS(VirtualFile file,
+                        RomFSExtractionType type = RomFSExtractionType::Truncated);
+// Converts a VFS filesystem into a RomFS binary
+// Returns nullptr on failure
+VirtualFile CreateRomFS(VirtualDir dir);
 } // namespace FileSys
diff --git a/src/core/file_sys/romfs_factory.cpp b/src/core/file_sys/romfs_factory.cpp
index d9d90939e..d027a8d59 100644
--- a/src/core/file_sys/romfs_factory.cpp
+++ b/src/core/file_sys/romfs_factory.cpp
@@ -28,11 +28,13 @@ RomFSFactory::RomFSFactory(Loader::AppLoader& app_loader) {
    ivfc_offset = app_loader.ReadRomFSIVFCOffset();
 }
+RomFSFactory::~RomFSFactory() = default;
 ResultVal<VirtualFile> RomFSFactory::OpenCurrentProcess() {
    if (!updatable)
        return MakeResult<VirtualFile>(file);
-    const PatchManager patch_manager(Core::CurrentProcess()->program_id);
+    const PatchManager patch_manager(Core::CurrentProcess()->GetTitleID());
    return MakeResult<VirtualFile>(patch_manager.PatchRomFS(file, ivfc_offset));
 }
diff --git a/src/core/file_sys/romfs_factory.h b/src/core/file_sys/romfs_factory.h
index 26b8f46cc..2cace8180 100644
--- a/src/core/file_sys/romfs_factory.h
+++ b/src/core/file_sys/romfs_factory.h
@@ -30,6 +30,7 @@ enum class StorageId : u8 {
 class RomFSFactory {
 public:
    explicit RomFSFactory(Loader::AppLoader& app_loader);
+    ~RomFSFactory();
    ResultVal<VirtualFile> OpenCurrentProcess();
    ResultVal<VirtualFile> Open(u64 title_id, StorageId storage, ContentRecordType type);
diff --git a/src/core/file_sys/savedata_factory.cpp b/src/core/file_sys/savedata_factory.cpp
index e437d34e5..47f2ab9e0 100644
--- a/src/core/file_sys/savedata_factory.cpp
+++ b/src/core/file_sys/savedata_factory.cpp
@@ -20,6 +20,8 @@ std::string SaveDataDescriptor::DebugInfo() const {
 SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save_directory)) {}
+SaveDataFactory::~SaveDataFactory() = default;
 ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, SaveDataDescriptor meta) {
    if (meta.type == SaveDataType::SystemSaveData || meta.type == SaveDataType::SaveData) {
        if (meta.zero_1 != 0) {
@@ -79,16 +81,16 @@ std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType typ
    // According to switchbrew, if a save is of type SaveData and the title id field is 0, it should
    // be interpreted as the title id of the current process.
    if (type == SaveDataType::SaveData && title_id == 0)
-        title_id = Core::CurrentProcess()->program_id;
+        title_id = Core::CurrentProcess()->GetTitleID();
    std::string out;
    switch (space) {
    case SaveDataSpaceId::NandSystem:
-        out = "/system/save/";
+        out = "/system/";
        break;
    case SaveDataSpaceId::NandUser:
-        out = "/user/save/";
+        out = "/user/";
        break;
    default:
        ASSERT_MSG(false, "Unrecognized SaveDataSpaceId: {:02X}", static_cast<u8>(space));
@@ -96,9 +98,12 @@ std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType typ
    switch (type) {
    case SaveDataType::SystemSaveData:
-        return fmt::format("{}{:016X}/{:016X}{:016X}", out, save_id, user_id[1], user_id[0]);
+        return fmt::format("{}save/{:016X}/{:016X}{:016X}", out, save_id, user_id[1], user_id[0]);
    case SaveDataType::SaveData:
-        return fmt::format("{}{:016X}/{:016X}{:016X}/{:016X}", out, 0, user_id[1], user_id[0],
+        return fmt::format("{}save/{:016X}/{:016X}{:016X}/{:016X}", out, 0, user_id[1], user_id[0],
+                           title_id);
+    case SaveDataType::TemporaryStorage:
+        return fmt::format("{}temp/{:016X}/{:016X}{:016X}/{:016X}", out, 0, user_id[1], user_id[0],
                           title_id);
    default:
        ASSERT_MSG(false, "Unrecognized SaveDataType: {:02X}", static_cast<u8>(type));
diff --git a/src/core/file_sys/savedata_factory.h b/src/core/file_sys/savedata_factory.h
index ba978695b..d69ef6741 100644
--- a/src/core/file_sys/savedata_factory.h
+++ b/src/core/file_sys/savedata_factory.h
@@ -48,6 +48,7 @@ static_assert(sizeof(SaveDataDescriptor) == 0x40, "SaveDataDescriptor has incorr
 class SaveDataFactory {
 public:
    explicit SaveDataFactory(VirtualDir dir);
+    ~SaveDataFactory();
    ResultVal<VirtualDir> Open(SaveDataSpaceId space, SaveDataDescriptor meta);
diff --git a/src/core/file_sys/submission_package.h b/src/core/file_sys/submission_package.h
index 1120a4920..e85a2b76e 100644
--- a/src/core/file_sys/submission_package.h
+++ b/src/core/file_sys/submission_package.h
@@ -24,7 +24,7 @@ enum class ContentRecordType : u8;
 class NSP : public ReadOnlyVfsDirectory {
 public:
    explicit NSP(VirtualFile file);
-    ~NSP();
+    ~NSP() override;
    Loader::ResultStatus GetStatus() const;
    Loader::ResultStatus GetProgramStatus(u64 title_id) const;
diff --git a/src/core/file_sys/vfs.cpp b/src/core/file_sys/vfs.cpp
index d7b52abfd..bfe50da73 100644
--- a/src/core/file_sys/vfs.cpp
+++ b/src/core/file_sys/vfs.cpp
@@ -399,6 +399,15 @@ bool VfsDirectory::Copy(std::string_view src, std::string_view dest) {
    return f2->WriteBytes(f1->ReadAllBytes()) == f1->GetSize();
 }
+std::map<std::string, VfsEntryType, std::less<>> VfsDirectory::GetEntries() const {
+    std::map<std::string, VfsEntryType, std::less<>> out;
+    for (const auto& dir : GetSubdirectories())
+        out.emplace(dir->GetName(), VfsEntryType::Directory);
+    for (const auto& file : GetFiles())
+        out.emplace(file->GetName(), VfsEntryType::File);
+    return out;
+}
 std::string VfsDirectory::GetFullPath() const {
    if (IsRoot())
        return GetName();
@@ -454,13 +463,41 @@ bool DeepEquals(const VirtualFile& file1, const VirtualFile& file2, std::size_t
    return true;
 }
-bool VfsRawCopy(VirtualFile src, VirtualFile dest) {
+bool VfsRawCopy(const VirtualFile& src, const VirtualFile& dest, std::size_t block_size) {
-    if (src == nullptr || dest == nullptr)
+    if (src == nullptr || dest == nullptr || !src->IsReadable() || !dest->IsWritable())
        return false;
    if (!dest->Resize(src->GetSize()))
        return false;
-    std::vector<u8> data = src->ReadAllBytes();
-    return dest->WriteBytes(data, 0) == data.size();
+    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)
+            return false;
+    }
+    return true;
+}
+bool VfsRawCopyD(const VirtualDir& src, const VirtualDir& dest, std::size_t block_size) {
+    if (src == nullptr || dest == nullptr || !src->IsReadable() || !dest->IsWritable())
+        return false;
+    for (const auto& file : src->GetFiles()) {
+        const auto out = dest->CreateFile(file->GetName());
+        if (!VfsRawCopy(file, out, block_size))
+            return false;
+    }
+    for (const auto& dir : src->GetSubdirectories()) {
+        const auto out = dest->CreateSubdirectory(dir->GetName());
+        if (!VfsRawCopyD(dir, out, block_size))
+            return false;
+    }
+    return true;
 }
 VirtualDir GetOrCreateDirectoryRelative(const VirtualDir& rel, std::string_view path) {
diff --git a/src/core/file_sys/vfs.h b/src/core/file_sys/vfs.h
index 74489b452..270291631 100644
--- a/src/core/file_sys/vfs.h
+++ b/src/core/file_sys/vfs.h
@@ -4,6 +4,7 @@
 #pragma once
+#include <map>
 #include <memory>
 #include <string>
 #include <string_view>
@@ -265,6 +266,10 @@ public:
    // dest.
    virtual bool Copy(std::string_view src, std::string_view dest);
+    // Gets all of the entries directly in the directory (files and dirs), returning a map between
+    // 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
@@ -310,13 +315,19 @@ public:
    bool Rename(std::string_view name) override;
 };
-// Compare the two files, byte-for-byte, in increments specificed by block_size
+// Compare the two files, byte-for-byte, in increments specified by block_size
-bool DeepEquals(const VirtualFile& file1, const VirtualFile& file2, std::size_t block_size = 0x200);
+bool DeepEquals(const VirtualFile& file1, const VirtualFile& file2,
+                std::size_t block_size = 0x1000);
 // A method that copies the raw data between two different implementations of VirtualFile. If you
 // are using the same implementation, it is probably better to use the Copy method in the parent
 // directory of src/dest.
-bool VfsRawCopy(VirtualFile src, VirtualFile dest);
+bool VfsRawCopy(const VirtualFile& src, const VirtualFile& dest, std::size_t block_size = 0x1000);
+// A method that performs a similar function to VfsRawCopy above, but instead copies entire
+// directories. It suffers the same performance penalties as above and an implementation-specific
+// Copy should always be preferred.
+bool VfsRawCopyD(const VirtualDir& src, const VirtualDir& dest, std::size_t block_size = 0x1000);
 // Checks if the directory at path relative to rel exists. If it does, returns that. If it does not
 // it attempts to create it and returns the new dir or nullptr on failure.
diff --git a/src/core/file_sys/vfs_concat.cpp b/src/core/file_sys/vfs_concat.cpp
index 25a980cbb..16d801c0c 100644
--- a/src/core/file_sys/vfs_concat.cpp
+++ b/src/core/file_sys/vfs_concat.cpp
@@ -5,17 +5,22 @@
 #include <algorithm>
 #include <utility>
+#include "common/assert.h"
 #include "core/file_sys/vfs_concat.h"
+#include "core/file_sys/vfs_static.h"
 namespace FileSys {
-VirtualFile ConcatenateFiles(std::vector<VirtualFile> files, std::string name) {
+static bool VerifyConcatenationMapContinuity(const std::map<u64, VirtualFile>& map) {
-    if (files.empty())
+    const auto last_valid = --map.end();
-        return nullptr;
+    for (auto iter = map.begin(); iter != last_valid;) {
-    if (files.size() == 1)
+        const auto old = iter++;
-        return files[0];
+        if (old->first + old->second->GetSize() != iter->first) {
+            return false;
+        }
+    }
-    return std::shared_ptr<VfsFile>(new ConcatenatedVfsFile(std::move(files), std::move(name)));
+    return map.begin()->first == 0;
 }
 ConcatenatedVfsFile::ConcatenatedVfsFile(std::vector<VirtualFile> files_, std::string name)
@@ -27,6 +32,48 @@ ConcatenatedVfsFile::ConcatenatedVfsFile(std::vector<VirtualFile> files_, std::s
    }
 }
+ConcatenatedVfsFile::ConcatenatedVfsFile(std::map<u64, VirtualFile> files_, std::string name)
+    : files(std::move(files_)), name(std::move(name)) {
+    ASSERT(VerifyConcatenationMapContinuity(files));
+}
+ConcatenatedVfsFile::~ConcatenatedVfsFile() = default;
+VirtualFile ConcatenatedVfsFile::MakeConcatenatedFile(std::vector<VirtualFile> files,
+                                                      std::string name) {
+    if (files.empty())
+        return nullptr;
+    if (files.size() == 1)
+        return files[0];
+    return std::shared_ptr<VfsFile>(new ConcatenatedVfsFile(std::move(files), std::move(name)));
+}
+VirtualFile ConcatenatedVfsFile::MakeConcatenatedFile(u8 filler_byte,
+                                                      std::map<u64, VirtualFile> files,
+                                                      std::string name) {
+    if (files.empty())
+        return nullptr;
+    if (files.size() == 1)
+        return files.begin()->second;
+    const auto last_valid = --files.end();
+    for (auto iter = files.begin(); iter != last_valid;) {
+        const auto old = iter++;
+        if (old->first + old->second->GetSize() != iter->first) {
+            files.emplace(old->first + old->second->GetSize(),
+                          std::make_shared<StaticVfsFile>(filler_byte, iter->first - old->first -
+                                                                           old->second->GetSize()));
+        }
+    }
+    // Ensure the map starts at offset 0 (start of file), otherwise pad to fill.
+    if (files.begin()->first != 0)
+        files.emplace(0, std::make_shared<StaticVfsFile>(filler_byte, files.begin()->first));
+    return std::shared_ptr<VfsFile>(new ConcatenatedVfsFile(std::move(files), std::move(name)));
+}
 std::string ConcatenatedVfsFile::GetName() const {
    if (files.empty())
        return "";
@@ -60,7 +107,7 @@ bool ConcatenatedVfsFile::IsReadable() const {
 }
 std::size_t ConcatenatedVfsFile::Read(u8* data, std::size_t length, std::size_t offset) const {
-    auto entry = files.end();
+    auto entry = --files.end();
    for (auto iter = files.begin(); iter != files.end(); ++iter) {
        if (iter->first > offset) {
            entry = --iter;
@@ -68,20 +115,17 @@ std::size_t ConcatenatedVfsFile::Read(u8* data, std::size_t length, std::size_t
        }
    }
-    // Check if the entry should be the last one. The loop above will make it end().
+    if (entry->first + entry->second->GetSize() <= offset)
-    if (entry == files.end() && offset < files.rbegin()->first + files.rbegin()->second->GetSize())
-        --entry;
-    if (entry == files.end())
        return 0;
-    const auto remaining = entry->second->GetSize() + offset - entry->first;
+    const auto read_in =
-    if (length > remaining) {
+        std::min<u64>(entry->first + entry->second->GetSize() - offset, entry->second->GetSize());
-        return entry->second->Read(data, remaining, offset - entry->first) +
+    if (length > read_in) {
-               Read(data + remaining, length - remaining, offset + remaining);
+        return entry->second->Read(data, read_in, offset - entry->first) +
+               Read(data + read_in, length - read_in, offset + read_in);
    }
-    return entry->second->Read(data, length, offset - entry->first);
+    return entry->second->Read(data, std::min<u64>(read_in, length), offset - entry->first);
 }
 std::size_t ConcatenatedVfsFile::Write(const u8* data, std::size_t length, std::size_t offset) {
@@ -91,4 +135,5 @@ std::size_t ConcatenatedVfsFile::Write(const u8* data, std::size_t length, std::
 bool ConcatenatedVfsFile::Rename(std::string_view name) {
    return false;
 }
 } // namespace FileSys
diff --git a/src/core/file_sys/vfs_concat.h b/src/core/file_sys/vfs_concat.h
index 31775db7e..c90f9d5d1 100644
--- a/src/core/file_sys/vfs_concat.h
+++ b/src/core/file_sys/vfs_concat.h
@@ -4,24 +4,30 @@
 #pragma once
+#include <map>
 #include <memory>
 #include <string_view>
-#include <boost/container/flat_map.hpp>
 #include "core/file_sys/vfs.h"
 namespace FileSys {
-// Wrapper function to allow for more efficient handling of files.size() == 0, 1 cases.
-VirtualFile ConcatenateFiles(std::vector<VirtualFile> files, std::string name = "");
 // Class that wraps multiple vfs files and concatenates them, making reads seamless. Currently
 // read-only.
 class ConcatenatedVfsFile : public VfsFile {
-    friend VirtualFile ConcatenateFiles(std::vector<VirtualFile> files, std::string name);
    ConcatenatedVfsFile(std::vector<VirtualFile> files, std::string name);
+    ConcatenatedVfsFile(std::map<u64, VirtualFile> files, std::string name);
 public:
+    ~ConcatenatedVfsFile() override;
+    /// Wrapper function to allow for more efficient handling of files.size() == 0, 1 cases.
+    static VirtualFile MakeConcatenatedFile(std::vector<VirtualFile> files, std::string name);
+    /// Convenience function that turns a map of offsets to files into a concatenated file, filling
+    /// gaps with a given filler byte.
+    static VirtualFile MakeConcatenatedFile(u8 filler_byte, std::map<u64, VirtualFile> files,
+                                            std::string name);
    std::string GetName() const override;
    std::size_t GetSize() const override;
    bool Resize(std::size_t new_size) override;
@@ -34,7 +40,7 @@ public:
 private:
    // Maps starting offset to file -- more efficient.
-    boost::container::flat_map<u64, VirtualFile> files;
+    std::map<u64, VirtualFile> files;
    std::string name;
 };
diff --git a/src/core/file_sys/vfs_layered.cpp b/src/core/file_sys/vfs_layered.cpp
new file mode 100644
index 000000000..bfee01725
--- /dev/null
+++ b/src/core/file_sys/vfs_layered.cpp
@@ -0,0 +1,132 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+#include <algorithm>
+#include <utility>
+#include "core/file_sys/vfs_layered.h"
+namespace FileSys {
+LayeredVfsDirectory::LayeredVfsDirectory(std::vector<VirtualDir> dirs, std::string name)
+    : dirs(std::move(dirs)), name(std::move(name)) {}
+LayeredVfsDirectory::~LayeredVfsDirectory() = default;
+VirtualDir LayeredVfsDirectory::MakeLayeredDirectory(std::vector<VirtualDir> dirs,
+                                                     std::string name) {
+    if (dirs.empty())
+        return nullptr;
+    if (dirs.size() == 1)
+        return dirs[0];
+    return std::shared_ptr<VfsDirectory>(new LayeredVfsDirectory(std::move(dirs), std::move(name)));
+}
+std::shared_ptr<VfsFile> LayeredVfsDirectory::GetFileRelative(std::string_view path) const {
+    for (const auto& layer : dirs) {
+        const auto file = layer->GetFileRelative(path);
+        if (file != nullptr)
+            return file;
+    }
+    return nullptr;
+}
+std::shared_ptr<VfsDirectory> LayeredVfsDirectory::GetDirectoryRelative(
+    std::string_view path) const {
+    std::vector<VirtualDir> out;
+    for (const auto& layer : dirs) {
+        auto dir = layer->GetDirectoryRelative(path);
+        if (dir != nullptr)
+            out.push_back(std::move(dir));
+    }
+    return MakeLayeredDirectory(std::move(out));
+}
+std::shared_ptr<VfsFile> LayeredVfsDirectory::GetFile(std::string_view name) const {
+    return GetFileRelative(name);
+}
+std::shared_ptr<VfsDirectory> LayeredVfsDirectory::GetSubdirectory(std::string_view name) const {
+    return GetDirectoryRelative(name);
+}
+std::string LayeredVfsDirectory::GetFullPath() const {
+    return dirs[0]->GetFullPath();
+}
+std::vector<std::shared_ptr<VfsFile>> LayeredVfsDirectory::GetFiles() const {
+    std::vector<VirtualFile> out;
+    for (const auto& layer : dirs) {
+        for (const auto& file : layer->GetFiles()) {
+            if (std::find_if(out.begin(), out.end(), [&file](const VirtualFile& comp) {
+                    return comp->GetName() == file->GetName();
+                }) == out.end()) {
+                out.push_back(file);
+            }
+        }
+    }
+    return out;
+}
+std::vector<std::shared_ptr<VfsDirectory>> LayeredVfsDirectory::GetSubdirectories() const {
+    std::vector<std::string> names;
+    for (const auto& layer : dirs) {
+        for (const auto& sd : layer->GetSubdirectories()) {
+            if (std::find(names.begin(), names.end(), sd->GetName()) == names.end())
+                names.push_back(sd->GetName());
+        }
+    }
+    std::vector<VirtualDir> out;
+    out.reserve(names.size());
+    for (const auto& subdir : names)
+        out.push_back(GetSubdirectory(subdir));
+    return out;
+}
+bool LayeredVfsDirectory::IsWritable() const {
+    return false;
+}
+bool LayeredVfsDirectory::IsReadable() const {
+    return true;
+}
+std::string LayeredVfsDirectory::GetName() const {
+    return name.empty() ? dirs[0]->GetName() : name;
+}
+std::shared_ptr<VfsDirectory> LayeredVfsDirectory::GetParentDirectory() const {
+    return dirs[0]->GetParentDirectory();
+}
+std::shared_ptr<VfsDirectory> LayeredVfsDirectory::CreateSubdirectory(std::string_view name) {
+    return nullptr;
+}
+std::shared_ptr<VfsFile> LayeredVfsDirectory::CreateFile(std::string_view name) {
+    return nullptr;
+}
+bool LayeredVfsDirectory::DeleteSubdirectory(std::string_view name) {
+    return false;
+}
+bool LayeredVfsDirectory::DeleteFile(std::string_view name) {
+    return false;
+}
+bool LayeredVfsDirectory::Rename(std::string_view name_) {
+    name = name_;
+    return true;
+}
+bool LayeredVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
+    return false;
+}
+} // namespace FileSys
diff --git a/src/core/file_sys/vfs_layered.h b/src/core/file_sys/vfs_layered.h
new file mode 100644
index 000000000..d85310f57
--- /dev/null
+++ b/src/core/file_sys/vfs_layered.h
@@ -0,0 +1,50 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+#pragma once
+#include <memory>
+#include "core/file_sys/vfs.h"
+namespace FileSys {
+// Class that stacks multiple VfsDirectories on top of each other, attempting to read from the first
+// one and falling back to the one after. The highest priority directory (overwrites all others)
+// should be element 0 in the dirs vector.
+class LayeredVfsDirectory : public VfsDirectory {
+    LayeredVfsDirectory(std::vector<VirtualDir> dirs, std::string name);
+public:
+    ~LayeredVfsDirectory() override;
+    /// Wrapper function to allow for more efficient handling of dirs.size() == 0, 1 cases.
+    static VirtualDir MakeLayeredDirectory(std::vector<VirtualDir> dirs, std::string name = "");
+    std::shared_ptr<VfsFile> GetFileRelative(std::string_view path) const override;
+    std::shared_ptr<VfsDirectory> GetDirectoryRelative(std::string_view path) const override;
+    std::shared_ptr<VfsFile> GetFile(std::string_view name) const override;
+    std::shared_ptr<VfsDirectory> GetSubdirectory(std::string_view name) const override;
+    std::string GetFullPath() const override;
+    std::vector<std::shared_ptr<VfsFile>> GetFiles() const override;
+    std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override;
+    bool IsWritable() const override;
+    bool IsReadable() const override;
+    std::string GetName() const override;
+    std::shared_ptr<VfsDirectory> GetParentDirectory() const override;
+    std::shared_ptr<VfsDirectory> CreateSubdirectory(std::string_view name) override;
+    std::shared_ptr<VfsFile> CreateFile(std::string_view name) override;
+    bool DeleteSubdirectory(std::string_view name) override;
+    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;
+};
+} // namespace FileSys
diff --git a/src/core/file_sys/vfs_offset.cpp b/src/core/file_sys/vfs_offset.cpp
index f5ed291ea..a4c6719a0 100644
--- a/src/core/file_sys/vfs_offset.cpp
+++ b/src/core/file_sys/vfs_offset.cpp
@@ -14,6 +14,8 @@ OffsetVfsFile::OffsetVfsFile(std::shared_ptr<VfsFile> file_, std::size_t size_,
    : file(file_), offset(offset_), size(size_), name(std::move(name_)),
      parent(parent_ == nullptr ? file->GetContainingDirectory() : std::move(parent_)) {}
+OffsetVfsFile::~OffsetVfsFile() = default;
 std::string OffsetVfsFile::GetName() const {
    return name.empty() ? file->GetName() : name;
 }
diff --git a/src/core/file_sys/vfs_offset.h b/src/core/file_sys/vfs_offset.h
index 34cb180b3..8062702a7 100644
--- a/src/core/file_sys/vfs_offset.h
+++ b/src/core/file_sys/vfs_offset.h
@@ -19,6 +19,7 @@ class OffsetVfsFile : public VfsFile {
 public:
    OffsetVfsFile(std::shared_ptr<VfsFile> file, std::size_t size, std::size_t offset = 0,
                  std::string new_name = "", VirtualDir new_parent = nullptr);
+    ~OffsetVfsFile() override;
    std::string GetName() const override;
    std::size_t GetSize() const override;
diff --git a/src/core/file_sys/vfs_real.cpp b/src/core/file_sys/vfs_real.cpp
index 5e242e20f..9defad04c 100644
--- a/src/core/file_sys/vfs_real.cpp
+++ b/src/core/file_sys/vfs_real.cpp
@@ -413,6 +413,23 @@ std::string RealVfsDirectory::GetFullPath() const {
    return out;
 }
+std::map<std::string, VfsEntryType, std::less<>> RealVfsDirectory::GetEntries() const {
+    if (perms == Mode::Append)
+        return {};
+    std::map<std::string, VfsEntryType, std::less<>> out;
+    FileUtil::ForeachDirectoryEntry(
+        nullptr, path,
+        [&out](u64* entries_out, const std::string& directory, const std::string& filename) {
+            const std::string full_path = directory + DIR_SEP + filename;
+            out.emplace(filename, FileUtil::IsDirectory(full_path) ? VfsEntryType::Directory
+                                                                   : VfsEntryType::File);
+            return true;
+        });
+    return out;
+}
 bool RealVfsDirectory::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) {
    return false;
 }
diff --git a/src/core/file_sys/vfs_real.h b/src/core/file_sys/vfs_real.h
index 681c43e82..5b61db90d 100644
--- a/src/core/file_sys/vfs_real.h
+++ b/src/core/file_sys/vfs_real.h
@@ -98,6 +98,7 @@ public:
    bool DeleteFile(std::string_view name) override;
    bool Rename(std::string_view name) override;
    std::string GetFullPath() const override;
+    std::map<std::string, VfsEntryType, std::less<>> GetEntries() const override;
 protected:
    bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override;
diff --git a/src/core/file_sys/vfs_static.h b/src/core/file_sys/vfs_static.h
new file mode 100644
index 000000000..44fab51d1
--- /dev/null
+++ b/src/core/file_sys/vfs_static.h
@@ -0,0 +1,79 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+#pragma once
+#include <algorithm>
+#include <memory>
+#include <string_view>
+#include "core/file_sys/vfs.h"
+namespace FileSys {
+class StaticVfsFile : public VfsFile {
+public:
+    explicit StaticVfsFile(u8 value, std::size_t size = 0, std::string name = "",
+                           VirtualDir parent = nullptr)
+        : value{value}, size{size}, name{std::move(name)}, parent{std::move(parent)} {}
+    std::string GetName() const override {
+        return name;
+    }
+    std::size_t GetSize() const override {
+        return size;
+    }
+    bool Resize(std::size_t new_size) override {
+        size = new_size;
+        return true;
+    }
+    std::shared_ptr<VfsDirectory> GetContainingDirectory() const override {
+        return parent;
+    }
+    bool IsWritable() const override {
+        return false;
+    }
+    bool IsReadable() const override {
+        return true;
+    }
+    std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override {
+        const auto read = std::min(length, size - offset);
+        std::fill(data, data + read, value);
+        return read;
+    }
+    std::size_t Write(const u8* data, std::size_t length, std::size_t offset) override {
+        return 0;
+    }
+    boost::optional<u8> ReadByte(std::size_t offset) const override {
+        if (offset < size)
+            return value;
+        return boost::none;
+    }
+    std::vector<u8> ReadBytes(std::size_t length, std::size_t offset) const override {
+        const auto read = std::min(length, size - offset);
+        return std::vector<u8>(read, value);
+    }
+    bool Rename(std::string_view new_name) override {
+        name = new_name;
+        return true;
+    }
+private:
+    u8 value;
+    std::size_t size;
+    std::string name;
+    VirtualDir parent;
+};
+} // namespace FileSys
diff --git a/src/core/file_sys/vfs_vector.cpp b/src/core/file_sys/vfs_vector.cpp
index 98e7c4598..389c7e003 100644
--- a/src/core/file_sys/vfs_vector.cpp
+++ b/src/core/file_sys/vfs_vector.cpp
@@ -3,16 +3,72 @@
 // Refer to the license.txt file included.
 #include <algorithm>
+#include <cstring>
 #include <utility>
 #include "core/file_sys/vfs_vector.h"
 namespace FileSys {
+VectorVfsFile::VectorVfsFile(std::vector<u8> initial_data, std::string name, VirtualDir parent)
+    : data(std::move(initial_data)), parent(std::move(parent)), name(std::move(name)) {}
+VectorVfsFile::~VectorVfsFile() = default;
+std::string VectorVfsFile::GetName() const {
+    return name;
+}
+size_t VectorVfsFile::GetSize() const {
+    return data.size();
+}
+bool VectorVfsFile::Resize(size_t new_size) {
+    data.resize(new_size);
+    return true;
+}
+std::shared_ptr<VfsDirectory> VectorVfsFile::GetContainingDirectory() const {
+    return parent;
+}
+bool VectorVfsFile::IsWritable() const {
+    return true;
+}
+bool VectorVfsFile::IsReadable() const {
+    return true;
+}
+std::size_t VectorVfsFile::Read(u8* data_, std::size_t length, std::size_t offset) const {
+    const auto read = std::min(length, data.size() - offset);
+    std::memcpy(data_, data.data() + offset, read);
+    return read;
+}
+std::size_t VectorVfsFile::Write(const u8* data_, std::size_t length, std::size_t offset) {
+    if (offset + length > data.size())
+        data.resize(offset + length);
+    const auto write = std::min(length, data.size() - offset);
+    std::memcpy(data.data(), data_, write);
+    return write;
+}
+bool VectorVfsFile::Rename(std::string_view name_) {
+    name = name_;
+    return true;
+}
+void VectorVfsFile::Assign(std::vector<u8> new_data) {
+    data = std::move(new_data);
+}
 VectorVfsDirectory::VectorVfsDirectory(std::vector<VirtualFile> files_,
                                       std::vector<VirtualDir> dirs_, std::string name_,
                                       VirtualDir parent_)
    : files(std::move(files_)), dirs(std::move(dirs_)), parent(std::move(parent_)),
      name(std::move(name_)) {}
+VectorVfsDirectory::~VectorVfsDirectory() = default;
 std::vector<std::shared_ptr<VfsFile>> VectorVfsDirectory::GetFiles() const {
    return files;
 }
diff --git a/src/core/file_sys/vfs_vector.h b/src/core/file_sys/vfs_vector.h
index 179f62e4b..48a414c98 100644
--- a/src/core/file_sys/vfs_vector.h
+++ b/src/core/file_sys/vfs_vector.h
@@ -8,6 +8,31 @@
 namespace FileSys {
+// An implementation of VfsFile that is backed by a vector optionally supplied upon construction
+class VectorVfsFile : public VfsFile {
+public:
+    explicit VectorVfsFile(std::vector<u8> initial_data = {}, std::string name = "",
+                           VirtualDir parent = nullptr);
+    ~VectorVfsFile() override;
+    std::string GetName() const override;
+    std::size_t GetSize() const override;
+    bool Resize(std::size_t new_size) override;
+    std::shared_ptr<VfsDirectory> GetContainingDirectory() const override;
+    bool IsWritable() const override;
+    bool IsReadable() const override;
+    std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override;
+    std::size_t Write(const u8* data, std::size_t length, std::size_t offset) override;
+    bool Rename(std::string_view name) override;
+    virtual void Assign(std::vector<u8> new_data);
+private:
+    std::vector<u8> data;
+    VirtualDir parent;
+    std::string name;
+};
 // An implementation of VfsDirectory that maintains two vectors for subdirectories and files.
 // Vector data is supplied upon construction.
 class VectorVfsDirectory : public VfsDirectory {
@@ -15,6 +40,7 @@ public:
    explicit VectorVfsDirectory(std::vector<VirtualFile> files = {},
                                std::vector<VirtualDir> dirs = {}, std::string name = "",
                                VirtualDir parent = nullptr);
+    ~VectorVfsDirectory() override;
    std::vector<std::shared_ptr<VfsFile>> GetFiles() const override;
    std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override;
diff --git a/src/core/file_sys/xts_archive.cpp b/src/core/file_sys/xts_archive.cpp
index 0173f71c1..b2b164368 100644
--- a/src/core/file_sys/xts_archive.cpp
+++ b/src/core/file_sys/xts_archive.cpp
@@ -30,9 +30,6 @@ static bool CalculateHMAC256(Destination* out, const SourceKey* key, std::size_t
    mbedtls_md_context_t context;
    mbedtls_md_init(&context);
-    const auto key_f = reinterpret_cast<const u8*>(key);
-    const std::vector<u8> key_v(key_f, key_f + key_length);
    if (mbedtls_md_setup(&context, mbedtls_md_info_from_type(MBEDTLS_MD_SHA256), 1) ||
        mbedtls_md_hmac_starts(&context, reinterpret_cast<const u8*>(key), key_length) ||
        mbedtls_md_hmac_update(&context, reinterpret_cast<const u8*>(data), data_length) ||
@@ -45,7 +42,7 @@ static bool CalculateHMAC256(Destination* out, const SourceKey* key, std::size_t
    return true;
 }
-NAX::NAX(VirtualFile file_) : file(std::move(file_)), header(std::make_unique<NAXHeader>()) {
+NAX::NAX(VirtualFile file_) : header(std::make_unique<NAXHeader>()), file(std::move(file_)) {
    std::string path = FileUtil::SanitizePath(file->GetFullPath());
    static const std::regex nax_path_regex("/registered/(000000[0-9A-F]{2})/([0-9A-F]{32})\\.nca",
                                           std::regex_constants::ECMAScript |
@@ -65,13 +62,15 @@ NAX::NAX(VirtualFile file_) : file(std::move(file_)), header(std::make_unique<NA
 }
 NAX::NAX(VirtualFile file_, std::array<u8, 0x10> nca_id)
-    : file(std::move(file_)), header(std::make_unique<NAXHeader>()) {
+    : header(std::make_unique<NAXHeader>()), file(std::move(file_)) {
    Core::Crypto::SHA256Hash hash{};
    mbedtls_sha256(nca_id.data(), nca_id.size(), hash.data(), 0);
    status = Parse(fmt::format("/registered/000000{:02X}/{}.nca", hash[0],
                               Common::HexArrayToString(nca_id, false)));
 }
+NAX::~NAX() = default;
 Loader::ResultStatus NAX::Parse(std::string_view path) {
    if (file->ReadObject(header.get()) != sizeof(NAXHeader))
        return Loader::ResultStatus::ErrorBadNAXHeader;
@@ -138,9 +137,9 @@ VirtualFile NAX::GetDecrypted() const {
    return dec_file;
 }
-std::shared_ptr<NCA> NAX::AsNCA() const {
+std::unique_ptr<NCA> NAX::AsNCA() const {
    if (type == NAXContentType::NCA)
-        return std::make_shared<NCA>(GetDecrypted());
+        return std::make_unique<NCA>(GetDecrypted());
    return nullptr;
 }
diff --git a/src/core/file_sys/xts_archive.h b/src/core/file_sys/xts_archive.h
index 55d2154a6..8fedd8585 100644
--- a/src/core/file_sys/xts_archive.h
+++ b/src/core/file_sys/xts_archive.h
@@ -33,12 +33,13 @@ class NAX : public ReadOnlyVfsDirectory {
 public:
    explicit NAX(VirtualFile file);
    explicit NAX(VirtualFile file, std::array<u8, 0x10> nca_id);
+    ~NAX() override;
    Loader::ResultStatus GetStatus() const;
    VirtualFile GetDecrypted() const;
-    std::shared_ptr<NCA> AsNCA() const;
+    std::unique_ptr<NCA> AsNCA() const;
    NAXContentType GetContentType() const;
@@ -60,7 +61,7 @@ private:
    VirtualFile file;
    Loader::ResultStatus status;
-    NAXContentType type;
+    NAXContentType type{};
    VirtualFile dec_file;
diff --git a/src/core/gdbstub/gdbstub.cpp b/src/core/gdbstub/gdbstub.cpp
index 1b04f68bf..5bc947010 100644
--- a/src/core/gdbstub/gdbstub.cpp
+++ b/src/core/gdbstub/gdbstub.cpp
@@ -37,7 +37,9 @@
 #include "core/core.h"
 #include "core/core_cpu.h"
 #include "core/gdbstub/gdbstub.h"
+#include "core/hle/kernel/process.h"
 #include "core/hle/kernel/scheduler.h"
+#include "core/hle/kernel/vm_manager.h"
 #include "core/loader/loader.h"
 #include "core/memory.h"
@@ -248,7 +250,7 @@ static void RegWrite(std::size_t id, u64 val, Kernel::Thread* thread = nullptr)
    } else if (id == PC_REGISTER) {
        thread->context.pc = val;
    } else if (id == PSTATE_REGISTER) {
-        thread->context.pstate = val;
+        thread->context.pstate = static_cast<u32>(val);
    } else if (id > PSTATE_REGISTER && id < FPCR_REGISTER) {
        thread->context.vector_registers[id - (PSTATE_REGISTER + 1)][0] = val;
    }
@@ -585,7 +587,8 @@ static void HandleQuery() {
                       strlen("Xfer:features:read:target.xml:")) == 0) {
        SendReply(target_xml);
    } else if (strncmp(query, "Offsets", strlen("Offsets")) == 0) {
-        std::string buffer = fmt::format("TextSeg={:0x}", Memory::PROCESS_IMAGE_VADDR);
+        const VAddr base_address = Core::CurrentProcess()->VMManager().GetCodeRegionBaseAddress();
+        std::string buffer = fmt::format("TextSeg={:0x}", base_address);
        SendReply(buffer.c_str());
    } else if (strncmp(query, "fThreadInfo", strlen("fThreadInfo")) == 0) {
        std::string val = "m";
@@ -893,11 +896,11 @@ static void ReadMemory() {
    static u8 reply[GDB_BUFFER_SIZE - 4];
    auto start_offset = command_buffer + 1;
-    auto addr_pos = std::find(start_offset, command_buffer + command_length, ',');
+    const auto addr_pos = std::find(start_offset, command_buffer + command_length, ',');
-    VAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset));
+    const VAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset));
    start_offset = addr_pos + 1;
-    u64 len =
+    const u64 len =
        HexToLong(start_offset, static_cast<u64>((command_buffer + command_length) - start_offset));
    LOG_DEBUG(Debug_GDBStub, "gdb: addr: {:016X} len: {:016X}", addr, len);
@@ -906,7 +909,9 @@ static void ReadMemory() {
        SendReply("E01");
    }
-    if (addr < Memory::PROCESS_IMAGE_VADDR || addr >= Memory::MAP_REGION_VADDR_END) {
+    const auto& vm_manager = Core::CurrentProcess()->VMManager();
+    if (addr < vm_manager.GetCodeRegionBaseAddress() ||
+        addr >= vm_manager.GetMapRegionEndAddress()) {
        return SendReply("E00");
    }
@@ -995,7 +1000,7 @@ static bool CommitBreakpoint(BreakpointType type, VAddr addr, u64 len) {
    breakpoint.addr = addr;
    breakpoint.len = len;
    Memory::ReadBlock(addr, breakpoint.inst.data(), breakpoint.inst.size());
-    static constexpr std::array<u8, 4> btrap{{0xd4, 0x20, 0x7d, 0x0}};
+    static constexpr std::array<u8, 4> btrap{{0x00, 0x7d, 0x20, 0xd4}};
    Memory::WriteBlock(addr, btrap.data(), btrap.size());
    Core::System::GetInstance().InvalidateCpuInstructionCaches();
    p.insert({addr, breakpoint});
diff --git a/src/core/hle/ipc_helpers.h b/src/core/hle/ipc_helpers.h
index 7545ecf2a..a4bfe2eb0 100644
--- a/src/core/hle/ipc_helpers.h
+++ b/src/core/hle/ipc_helpers.h
@@ -290,13 +290,6 @@ public:
        Skip(CommandIdSize, false);
    }
-    ResponseBuilder MakeBuilder(u32 normal_params_size, u32 num_handles_to_copy,
-                                u32 num_handles_to_move,
-                                ResponseBuilder::Flags flags = ResponseBuilder::Flags::None) const {
-        return ResponseBuilder{*context, normal_params_size, num_handles_to_copy,
-                               num_handles_to_move, flags};
-    }
    template <typename T>
    T Pop();
diff --git a/src/core/hle/kernel/errors.h b/src/core/hle/kernel/errors.h
index 8c2be2681..e5fa67ae8 100644
--- a/src/core/hle/kernel/errors.h
+++ b/src/core/hle/kernel/errors.h
@@ -31,6 +31,7 @@ enum {
    TooLarge = 119,
    InvalidEnumValue = 120,
    NoSuchEntry = 121,
+    AlreadyRegistered = 122,
    InvalidState = 125,
    ResourceLimitExceeded = 132,
 };
@@ -58,6 +59,7 @@ constexpr ResultCode ERR_INVALID_MEMORY_PERMISSIONS(ErrorModule::Kernel,
 constexpr ResultCode ERR_INVALID_HANDLE(ErrorModule::Kernel, ErrCodes::InvalidHandle);
 constexpr ResultCode ERR_INVALID_PROCESSOR_ID(ErrorModule::Kernel, ErrCodes::InvalidProcessorId);
 constexpr ResultCode ERR_INVALID_SIZE(ErrorModule::Kernel, ErrCodes::InvalidSize);
+constexpr ResultCode ERR_ALREADY_REGISTERED(ErrorModule::Kernel, ErrCodes::AlreadyRegistered);
 constexpr ResultCode ERR_INVALID_STATE(ErrorModule::Kernel, ErrCodes::InvalidState);
 constexpr ResultCode ERR_INVALID_THREAD_PRIORITY(ErrorModule::Kernel,
                                                 ErrCodes::InvalidThreadPriority);
diff --git a/src/core/hle/kernel/mutex.cpp b/src/core/hle/kernel/mutex.cpp
index 51f4544be..81675eac5 100644
--- a/src/core/hle/kernel/mutex.cpp
+++ b/src/core/hle/kernel/mutex.cpp
@@ -16,6 +16,7 @@
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/result.h"
+#include "core/memory.h"
 namespace Kernel {
diff --git a/src/core/hle/kernel/object.h b/src/core/hle/kernel/object.h
index b054cbf7d..9eb72315c 100644
--- a/src/core/hle/kernel/object.h
+++ b/src/core/hle/kernel/object.h
@@ -6,7 +6,6 @@
 #include <atomic>
 #include <string>
-#include <utility>
 #include <boost/smart_ptr/intrusive_ptr.hpp>
@@ -97,7 +96,7 @@ using SharedPtr = boost::intrusive_ptr<T>;
 template <typename T>
 inline SharedPtr<T> DynamicObjectCast(SharedPtr<Object> object) {
    if (object != nullptr && object->GetHandleType() == T::HANDLE_TYPE) {
-        return boost::static_pointer_cast<T>(std::move(object));
+        return boost::static_pointer_cast<T>(object);
    }
    return nullptr;
 }
diff --git a/src/core/hle/kernel/process.cpp b/src/core/hle/kernel/process.cpp
index 7a272d031..dc9fc8470 100644
--- a/src/core/hle/kernel/process.cpp
+++ b/src/core/hle/kernel/process.cpp
@@ -7,10 +7,13 @@
 #include "common/assert.h"
 #include "common/common_funcs.h"
 #include "common/logging/log.h"
+#include "core/core.h"
+#include "core/file_sys/program_metadata.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/resource_limit.h"
+#include "core/hle/kernel/scheduler.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/vm_manager.h"
 #include "core/memory.h"
@@ -32,14 +35,22 @@ SharedPtr<Process> Process::Create(KernelCore& kernel, std::string&& name) {
    process->name = std::move(name);
    process->flags.raw = 0;
    process->flags.memory_region.Assign(MemoryRegion::APPLICATION);
+    process->resource_limit = kernel.ResourceLimitForCategory(ResourceLimitCategory::APPLICATION);
    process->status = ProcessStatus::Created;
    process->program_id = 0;
    process->process_id = kernel.CreateNewProcessID();
+    process->svc_access_mask.set();
    kernel.AppendNewProcess(process);
    return process;
 }
+void Process::LoadFromMetadata(const FileSys::ProgramMetadata& metadata) {
+    program_id = metadata.GetTitleID();
+    is_64bit_process = metadata.Is64BitProgram();
+    vm_manager.Reset(metadata.GetAddressSpaceType());
+}
 void Process::ParseKernelCaps(const u32* kernel_caps, std::size_t len) {
    for (std::size_t i = 0; i < len; ++i) {
        u32 descriptor = kernel_caps[i];
@@ -117,7 +128,7 @@ void Process::Run(VAddr entry_point, s32 main_thread_priority, u32 stack_size) {
    // TODO(bunnei): This is heap area that should be allocated by the kernel and not mapped as part
    // of the user address space.
    vm_manager
-        .MapMemoryBlock(Memory::STACK_AREA_VADDR_END - stack_size,
+        .MapMemoryBlock(vm_manager.GetTLSIORegionEndAddress() - stack_size,
                        std::make_shared<std::vector<u8>>(stack_size, 0), 0, stack_size,
                        MemoryState::Mapped)
        .Unwrap();
@@ -125,7 +136,92 @@ void Process::Run(VAddr entry_point, s32 main_thread_priority, u32 stack_size) {
    vm_manager.LogLayout();
    status = ProcessStatus::Running;
-    Kernel::SetupMainThread(kernel, entry_point, main_thread_priority, this);
+    Kernel::SetupMainThread(kernel, entry_point, main_thread_priority, *this);
+}
+void Process::PrepareForTermination() {
+    status = ProcessStatus::Exited;
+    const auto stop_threads = [this](const std::vector<SharedPtr<Thread>>& thread_list) {
+        for (auto& thread : thread_list) {
+            if (thread->owner_process != this)
+                continue;
+            if (thread == GetCurrentThread())
+                continue;
+            // TODO(Subv): When are the other running/ready threads terminated?
+            ASSERT_MSG(thread->status == ThreadStatus::WaitSynchAny ||
+                           thread->status == ThreadStatus::WaitSynchAll,
+                       "Exiting processes with non-waiting threads is currently unimplemented");
+            thread->Stop();
+        }
+    };
+    auto& system = Core::System::GetInstance();
+    stop_threads(system.Scheduler(0)->GetThreadList());
+    stop_threads(system.Scheduler(1)->GetThreadList());
+    stop_threads(system.Scheduler(2)->GetThreadList());
+    stop_threads(system.Scheduler(3)->GetThreadList());
+}
+/**
+ * Finds a free location for the TLS section of a thread.
+ * @param tls_slots The TLS page array of the thread's owner process.
+ * Returns a tuple of (page, slot, alloc_needed) where:
+ * page: The index of the first allocated TLS page that has free slots.
+ * slot: The index of the first free slot in the indicated page.
+ * alloc_needed: Whether there's a need to allocate a new TLS page (All pages are full).
+ */
+static std::tuple<std::size_t, std::size_t, bool> FindFreeThreadLocalSlot(
+    const std::vector<std::bitset<8>>& tls_slots) {
+    // Iterate over all the allocated pages, and try to find one where not all slots are used.
+    for (std::size_t page = 0; page < tls_slots.size(); ++page) {
+        const auto& page_tls_slots = tls_slots[page];
+        if (!page_tls_slots.all()) {
+            // We found a page with at least one free slot, find which slot it is
+            for (std::size_t slot = 0; slot < page_tls_slots.size(); ++slot) {
+                if (!page_tls_slots.test(slot)) {
+                    return std::make_tuple(page, slot, false);
+                }
+            }
+        }
+    }
+    return std::make_tuple(0, 0, true);
+}
+VAddr Process::MarkNextAvailableTLSSlotAsUsed(Thread& thread) {
+    auto [available_page, available_slot, needs_allocation] = FindFreeThreadLocalSlot(tls_slots);
+    const VAddr tls_begin = vm_manager.GetTLSIORegionBaseAddress();
+    if (needs_allocation) {
+        tls_slots.emplace_back(0); // The page is completely available at the start
+        available_page = tls_slots.size() - 1;
+        available_slot = 0; // Use the first slot in the new page
+        // Allocate some memory from the end of the linear heap for this region.
+        auto& tls_memory = thread.GetTLSMemory();
+        tls_memory->insert(tls_memory->end(), Memory::PAGE_SIZE, 0);
+        vm_manager.RefreshMemoryBlockMappings(tls_memory.get());
+        vm_manager.MapMemoryBlock(tls_begin + available_page * Memory::PAGE_SIZE, tls_memory, 0,
+                                  Memory::PAGE_SIZE, MemoryState::ThreadLocal);
+    }
+    tls_slots[available_page].set(available_slot);
+    return tls_begin + available_page * Memory::PAGE_SIZE + available_slot * Memory::TLS_ENTRY_SIZE;
+}
+void Process::FreeTLSSlot(VAddr tls_address) {
+    const VAddr tls_base = tls_address - vm_manager.GetTLSIORegionBaseAddress();
+    const VAddr tls_page = tls_base / Memory::PAGE_SIZE;
+    const VAddr tls_slot = (tls_base % Memory::PAGE_SIZE) / Memory::TLS_ENTRY_SIZE;
+    tls_slots[tls_page].reset(tls_slot);
 }
 void Process::LoadModule(SharedPtr<CodeSet> module_, VAddr base_addr) {
@@ -145,8 +241,8 @@ void Process::LoadModule(SharedPtr<CodeSet> module_, VAddr base_addr) {
 }
 ResultVal<VAddr> Process::HeapAllocate(VAddr target, u64 size, VMAPermission perms) {
-    if (target < Memory::HEAP_VADDR || target + size > Memory::HEAP_VADDR_END ||
+    if (target < vm_manager.GetHeapRegionBaseAddress() ||
-        target + size < target) {
+        target + size > vm_manager.GetHeapRegionEndAddress() || target + size < target) {
        return ERR_INVALID_ADDRESS;
    }
@@ -181,8 +277,8 @@ ResultVal<VAddr> Process::HeapAllocate(VAddr target, u64 size, VMAPermission per
 }
 ResultCode Process::HeapFree(VAddr target, u32 size) {
-    if (target < Memory::HEAP_VADDR || target + size > Memory::HEAP_VADDR_END ||
+    if (target < vm_manager.GetHeapRegionBaseAddress() ||
-        target + size < target) {
+        target + size > vm_manager.GetHeapRegionEndAddress() || target + size < target) {
        return ERR_INVALID_ADDRESS;
    }
diff --git a/src/core/hle/kernel/process.h b/src/core/hle/kernel/process.h
index 81538f70c..590e0c73d 100644
--- a/src/core/hle/kernel/process.h
+++ b/src/core/hle/kernel/process.h
@@ -17,6 +17,10 @@
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/vm_manager.h"
+namespace FileSys {
+class ProgramMetadata;
+}
 namespace Kernel {
 class KernelCore;
@@ -131,6 +135,121 @@ public:
        return HANDLE_TYPE;
    }
+    /// Gets a reference to the process' memory manager.
+    Kernel::VMManager& VMManager() {
+        return vm_manager;
+    }
+    /// Gets a const reference to the process' memory manager.
+    const Kernel::VMManager& VMManager() const {
+        return vm_manager;
+    }
+    /// Gets the current status of the process
+    ProcessStatus GetStatus() const {
+        return status;
+    }
+    /// Gets the unique ID that identifies this particular process.
+    u32 GetProcessID() const {
+        return process_id;
+    }
+    /// Gets the title ID corresponding to this process.
+    u64 GetTitleID() const {
+        return program_id;
+    }
+    /// Gets the resource limit descriptor for this process
+    ResourceLimit& GetResourceLimit() {
+        return *resource_limit;
+    }
+    /// Gets the resource limit descriptor for this process
+    const ResourceLimit& GetResourceLimit() const {
+        return *resource_limit;
+    }
+    /// Gets the default CPU ID for this process
+    u8 GetDefaultProcessorID() const {
+        return ideal_processor;
+    }
+    /// Gets the bitmask of allowed CPUs that this process' threads can run on.
+    u32 GetAllowedProcessorMask() const {
+        return allowed_processor_mask;
+    }
+    /// Gets the bitmask of allowed thread priorities.
+    u32 GetAllowedThreadPriorityMask() const {
+        return allowed_thread_priority_mask;
+    }
+    u32 IsVirtualMemoryEnabled() const {
+        return is_virtual_address_memory_enabled;
+    }
+    /// Whether this process is an AArch64 or AArch32 process.
+    bool Is64BitProcess() const {
+        return is_64bit_process;
+    }
+    /**
+     * Loads process-specifics configuration info with metadata provided
+     * by an executable.
+     *
+     * @param metadata The provided metadata to load process specific info.
+     */
+    void LoadFromMetadata(const FileSys::ProgramMetadata& metadata);
+    /**
+     * Parses a list of kernel capability descriptors (as found in the ExHeader) and applies them
+     * to this process.
+     */
+    void ParseKernelCaps(const u32* kernel_caps, std::size_t len);
+    /**
+     * Applies address space changes and launches the process main thread.
+     */
+    void Run(VAddr entry_point, s32 main_thread_priority, u32 stack_size);
+    /**
+     * Prepares a process for termination by stopping all of its threads
+     * and clearing any other resources.
+     */
+    void PrepareForTermination();
+    void LoadModule(SharedPtr<CodeSet> module_, VAddr base_addr);
+    ///////////////////////////////////////////////////////////////////////////////////////////////
+    // Memory Management
+    // Marks the next available region as used and returns the address of the slot.
+    VAddr MarkNextAvailableTLSSlotAsUsed(Thread& thread);
+    // Frees a used TLS slot identified by the given address
+    void FreeTLSSlot(VAddr tls_address);
+    ResultVal<VAddr> HeapAllocate(VAddr target, u64 size, VMAPermission perms);
+    ResultCode HeapFree(VAddr target, u32 size);
+    ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size);
+    ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, u64 size);
+private:
+    explicit Process(KernelCore& kernel);
+    ~Process() override;
+    /// Memory manager for this process.
+    Kernel::VMManager vm_manager;
+    /// Current status of the process
+    ProcessStatus status;
+    /// The ID of this process
+    u32 process_id = 0;
    /// Title ID corresponding to the process
    u64 program_id;
@@ -140,7 +259,7 @@ public:
    /// The process may only call SVCs which have the corresponding bit set.
    std::bitset<0x80> svc_access_mask;
    /// Maximum size of the handle table for the process.
-    unsigned int handle_table_size = 0x200;
+    u32 handle_table_size = 0x200;
    /// Special memory ranges mapped into this processes address space. This is used to give
    /// processes access to specific I/O regions and device memory.
    boost::container::static_vector<AddressMapping, 8> address_mappings;
@@ -154,29 +273,6 @@ public:
    u32 allowed_processor_mask = THREADPROCESSORID_DEFAULT_MASK;
    u32 allowed_thread_priority_mask = 0xFFFFFFFF;
    u32 is_virtual_address_memory_enabled = 0;
-    /// Current status of the process
-    ProcessStatus status;
-    /// The ID of this process
-    u32 process_id = 0;
-    /**
-     * Parses a list of kernel capability descriptors (as found in the ExHeader) and applies them
-     * to this process.
-     */
-    void ParseKernelCaps(const u32* kernel_caps, std::size_t len);
-    /**
-     * Applies address space changes and launches the process main thread.
-     */
-    void Run(VAddr entry_point, s32 main_thread_priority, u32 stack_size);
-    void LoadModule(SharedPtr<CodeSet> module_, VAddr base_addr);
-    ///////////////////////////////////////////////////////////////////////////////////////////////
-    // Memory Management
-    VMManager vm_manager;
    // Memory used to back the allocations in the regular heap. A single vector is used to cover
    // the entire virtual address space extents that bound the allocations, including any holes.
@@ -196,18 +292,12 @@ public:
    /// This vector will grow as more pages are allocated for new threads.
    std::vector<std::bitset<8>> tls_slots;
-    std::string name;
+    /// Whether or not this process is AArch64, or AArch32.
+    /// By default, we currently assume this is true, unless otherwise
-    ResultVal<VAddr> HeapAllocate(VAddr target, u64 size, VMAPermission perms);
+    /// specified by metadata provided to the process during loading.
-    ResultCode HeapFree(VAddr target, u32 size);
+    bool is_64bit_process = true;
-    ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size);
-    ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, u64 size);
+    std::string name;
-private:
-    explicit Process(KernelCore& kernel);
-    ~Process() override;
 };
 } // namespace Kernel
diff --git a/src/core/hle/kernel/scheduler.cpp b/src/core/hle/kernel/scheduler.cpp
index 69c812f16..1e82cfffb 100644
--- a/src/core/hle/kernel/scheduler.cpp
+++ b/src/core/hle/kernel/scheduler.cpp
@@ -17,7 +17,7 @@ namespace Kernel {
 std::mutex Scheduler::scheduler_mutex;
-Scheduler::Scheduler(Core::ARM_Interface* cpu_core) : cpu_core(cpu_core) {}
+Scheduler::Scheduler(Core::ARM_Interface& cpu_core) : cpu_core(cpu_core) {}
 Scheduler::~Scheduler() {
    for (auto& thread : thread_list) {
@@ -59,9 +59,9 @@ void Scheduler::SwitchContext(Thread* new_thread) {
    // Save context for previous thread
    if (previous_thread) {
        previous_thread->last_running_ticks = CoreTiming::GetTicks();
-        cpu_core->SaveContext(previous_thread->context);
+        cpu_core.SaveContext(previous_thread->context);
        // Save the TPIDR_EL0 system register in case it was modified.
-        previous_thread->tpidr_el0 = cpu_core->GetTPIDR_EL0();
+        previous_thread->tpidr_el0 = cpu_core.GetTPIDR_EL0();
        if (previous_thread->status == ThreadStatus::Running) {
            // This is only the case when a reschedule is triggered without the current thread
@@ -88,13 +88,13 @@ void Scheduler::SwitchContext(Thread* new_thread) {
        if (previous_process != current_thread->owner_process) {
            Core::CurrentProcess() = current_thread->owner_process;
-            SetCurrentPageTable(&Core::CurrentProcess()->vm_manager.page_table);
+            SetCurrentPageTable(&Core::CurrentProcess()->VMManager().page_table);
        }
-        cpu_core->LoadContext(new_thread->context);
+        cpu_core.LoadContext(new_thread->context);
-        cpu_core->SetTlsAddress(new_thread->GetTLSAddress());
+        cpu_core.SetTlsAddress(new_thread->GetTLSAddress());
-        cpu_core->SetTPIDR_EL0(new_thread->GetTPIDR_EL0());
+        cpu_core.SetTPIDR_EL0(new_thread->GetTPIDR_EL0());
-        cpu_core->ClearExclusiveState();
+        cpu_core.ClearExclusiveState();
    } else {
        current_thread = nullptr;
        // Note: We do not reset the current process and current page table when idling because
diff --git a/src/core/hle/kernel/scheduler.h b/src/core/hle/kernel/scheduler.h
index 744990c9b..2c94641ec 100644
--- a/src/core/hle/kernel/scheduler.h
+++ b/src/core/hle/kernel/scheduler.h
@@ -19,7 +19,7 @@ namespace Kernel {
 class Scheduler final {
 public:
-    explicit Scheduler(Core::ARM_Interface* cpu_core);
+    explicit Scheduler(Core::ARM_Interface& cpu_core);
    ~Scheduler();
    /// Returns whether there are any threads that are ready to run.
@@ -72,7 +72,7 @@ private:
    SharedPtr<Thread> current_thread = nullptr;
-    Core::ARM_Interface* cpu_core;
+    Core::ARM_Interface& cpu_core;
    static std::mutex scheduler_mutex;
 };
diff --git a/src/core/hle/kernel/shared_memory.cpp b/src/core/hle/kernel/shared_memory.cpp
index abb1d09cd..d061e6155 100644
--- a/src/core/hle/kernel/shared_memory.cpp
+++ b/src/core/hle/kernel/shared_memory.cpp
@@ -8,6 +8,7 @@
 #include "common/logging/log.h"
 #include "core/core.h"
 #include "core/hle/kernel/errors.h"
+#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/shared_memory.h"
 #include "core/memory.h"
@@ -34,11 +35,11 @@ SharedPtr<SharedMemory> SharedMemory::Create(KernelCore& kernel, SharedPtr<Proce
        // Refresh the address mappings for the current process.
        if (Core::CurrentProcess() != nullptr) {
-            Core::CurrentProcess()->vm_manager.RefreshMemoryBlockMappings(
+            Core::CurrentProcess()->VMManager().RefreshMemoryBlockMappings(
                shared_memory->backing_block.get());
        }
    } else {
-        auto& vm_manager = shared_memory->owner_process->vm_manager;
+        auto& vm_manager = shared_memory->owner_process->VMManager();
        // The memory is already available and mapped in the owner process.
        auto vma = vm_manager.FindVMA(address);
@@ -71,7 +72,8 @@ SharedPtr<SharedMemory> SharedMemory::CreateForApplet(
    shared_memory->other_permissions = other_permissions;
    shared_memory->backing_block = std::move(heap_block);
    shared_memory->backing_block_offset = offset;
-    shared_memory->base_address = Memory::HEAP_VADDR + offset;
+    shared_memory->base_address =
+        kernel.CurrentProcess()->VMManager().GetHeapRegionBaseAddress() + offset;
    return shared_memory;
 }
@@ -105,7 +107,7 @@ ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermi
    VAddr target_address = address;
    // Map the memory block into the target process
-    auto result = target_process->vm_manager.MapMemoryBlock(
+    auto result = target_process->VMManager().MapMemoryBlock(
        target_address, backing_block, backing_block_offset, size, MemoryState::Shared);
    if (result.Failed()) {
        LOG_ERROR(
@@ -115,14 +117,14 @@ ResultCode SharedMemory::Map(Process* target_process, VAddr address, MemoryPermi
        return result.Code();
    }
-    return target_process->vm_manager.ReprotectRange(target_address, size,
+    return target_process->VMManager().ReprotectRange(target_address, size,
-                                                     ConvertPermissions(permissions));
+                                                      ConvertPermissions(permissions));
 }
 ResultCode SharedMemory::Unmap(Process* target_process, VAddr address) {
    // TODO(Subv): Verify what happens if the application tries to unmap an address that is not
    // mapped to a SharedMemory.
-    return target_process->vm_manager.UnmapRange(address, size);
+    return target_process->VMManager().UnmapRange(address, size);
 }
 VMAPermission SharedMemory::ConvertPermissions(MemoryPermission permission) {
diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp
index 371fc439e..1cdaa740a 100644
--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -51,8 +51,9 @@ static ResultCode SetHeapSize(VAddr* heap_addr, u64 heap_size) {
    }
    auto& process = *Core::CurrentProcess();
+    const VAddr heap_base = process.VMManager().GetHeapRegionBaseAddress();
    CASCADE_RESULT(*heap_addr,
-                   process.HeapAllocate(Memory::HEAP_VADDR, heap_size, VMAPermission::ReadWrite));
+                   process.HeapAllocate(heap_base, heap_size, VMAPermission::ReadWrite));
    return RESULT_SUCCESS;
 }
@@ -169,7 +170,7 @@ static ResultCode GetProcessId(u32* process_id, Handle process_handle) {
        return ERR_INVALID_HANDLE;
    }
-    *process_id = process->process_id;
+    *process_id = process->GetProcessID();
    return RESULT_SUCCESS;
 }
@@ -325,26 +326,27 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
    LOG_TRACE(Kernel_SVC, "called info_id=0x{:X}, info_sub_id=0x{:X}, handle=0x{:08X}", info_id,
              info_sub_id, handle);
-    const auto& vm_manager = Core::CurrentProcess()->vm_manager;
+    const auto& current_process = Core::CurrentProcess();
+    const auto& vm_manager = current_process->VMManager();
    switch (static_cast<GetInfoType>(info_id)) {
    case GetInfoType::AllowedCpuIdBitmask:
-        *result = Core::CurrentProcess()->allowed_processor_mask;
+        *result = current_process->GetAllowedProcessorMask();
        break;
    case GetInfoType::AllowedThreadPrioBitmask:
-        *result = Core::CurrentProcess()->allowed_thread_priority_mask;
+        *result = current_process->GetAllowedThreadPriorityMask();
        break;
    case GetInfoType::MapRegionBaseAddr:
-        *result = Memory::MAP_REGION_VADDR;
+        *result = vm_manager.GetMapRegionBaseAddress();
        break;
    case GetInfoType::MapRegionSize:
-        *result = Memory::MAP_REGION_SIZE;
+        *result = vm_manager.GetMapRegionSize();
        break;
    case GetInfoType::HeapRegionBaseAddr:
-        *result = Memory::HEAP_VADDR;
+        *result = vm_manager.GetHeapRegionBaseAddress();
        break;
    case GetInfoType::HeapRegionSize:
-        *result = Memory::HEAP_SIZE;
+        *result = vm_manager.GetHeapRegionSize();
        break;
    case GetInfoType::TotalMemoryUsage:
        *result = vm_manager.GetTotalMemoryUsage();
@@ -359,22 +361,35 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
        *result = 0;
        break;
    case GetInfoType::AddressSpaceBaseAddr:
-        *result = vm_manager.GetAddressSpaceBaseAddr();
+        *result = vm_manager.GetCodeRegionBaseAddress();
        break;
-    case GetInfoType::AddressSpaceSize:
+    case GetInfoType::AddressSpaceSize: {
-        *result = vm_manager.GetAddressSpaceSize();
+        const u64 width = vm_manager.GetAddressSpaceWidth();
+        switch (width) {
+        case 32:
+            *result = 0xFFE00000;
+            break;
+        case 36:
+            *result = 0xFF8000000;
+            break;
+        case 39:
+            *result = 0x7FF8000000;
+            break;
+        }
        break;
+    }
    case GetInfoType::NewMapRegionBaseAddr:
-        *result = Memory::NEW_MAP_REGION_VADDR;
+        *result = vm_manager.GetNewMapRegionBaseAddress();
        break;
    case GetInfoType::NewMapRegionSize:
-        *result = Memory::NEW_MAP_REGION_SIZE;
+        *result = vm_manager.GetNewMapRegionSize();
        break;
    case GetInfoType::IsVirtualAddressMemoryEnabled:
-        *result = Core::CurrentProcess()->is_virtual_address_memory_enabled;
+        *result = current_process->IsVirtualMemoryEnabled();
        break;
    case GetInfoType::TitleId:
-        *result = Core::CurrentProcess()->program_id;
+        *result = current_process->GetTitleID();
        break;
    case GetInfoType::PrivilegedProcessId:
        LOG_WARNING(Kernel_SVC,
@@ -400,8 +415,36 @@ static ResultCode SetThreadActivity(Handle handle, u32 unknown) {
 }
 /// Gets the thread context
-static ResultCode GetThreadContext(Handle handle, VAddr addr) {
+static ResultCode GetThreadContext(VAddr thread_context, Handle handle) {
-    LOG_WARNING(Kernel_SVC, "(STUBBED) called, handle=0x{:08X}, addr=0x{:X}", handle, addr);
+    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);
+    if (!thread) {
+        return ERR_INVALID_HANDLE;
+    }
+    const auto current_process = Core::CurrentProcess();
+    if (thread->owner_process != current_process) {
+        return ERR_INVALID_HANDLE;
+    }
+    if (thread == GetCurrentThread()) {
+        return ERR_ALREADY_REGISTERED;
+    }
+    Core::ARM_Interface::ThreadContext ctx = thread->context;
+    // Mask away mode bits, interrupt bits, IL bit, and other reserved bits.
+    ctx.pstate &= 0xFF0FFE20;
+    // If 64-bit, we can just write the context registers directly and we're good.
+    // However, if 32-bit, we have to ensure some registers are zeroed out.
+    if (!current_process->Is64BitProcess()) {
+        std::fill(ctx.cpu_registers.begin() + 15, ctx.cpu_registers.end(), 0);
+        std::fill(ctx.vector_registers.begin() + 16, ctx.vector_registers.end(), u128{});
+    }
+    Memory::WriteBlock(thread_context, &ctx, sizeof(ctx));
    return RESULT_SUCCESS;
 }
@@ -429,8 +472,8 @@ static ResultCode SetThreadPriority(Handle handle, u32 priority) {
    // Note: The kernel uses the current process's resource limit instead of
    // the one from the thread owner's resource limit.
-    SharedPtr<ResourceLimit>& resource_limit = Core::CurrentProcess()->resource_limit;
+    const ResourceLimit& resource_limit = Core::CurrentProcess()->GetResourceLimit();
-    if (resource_limit->GetMaxResourceValue(ResourceType::Priority) > priority) {
+    if (resource_limit.GetMaxResourceValue(ResourceType::Priority) > priority) {
        return ERR_NOT_AUTHORIZED;
    }
@@ -504,9 +547,9 @@ static ResultCode QueryProcessMemory(MemoryInfo* memory_info, PageInfo* /*page_i
    if (!process) {
        return ERR_INVALID_HANDLE;
    }
-    auto vma = process->vm_manager.FindVMA(addr);
+    auto vma = process->VMManager().FindVMA(addr);
    memory_info->attributes = 0;
-    if (vma == Core::CurrentProcess()->vm_manager.vma_map.end()) {
+    if (vma == Core::CurrentProcess()->VMManager().vma_map.end()) {
        memory_info->base_address = 0;
        memory_info->permission = static_cast<u32>(VMAPermission::None);
        memory_info->size = 0;
@@ -530,35 +573,13 @@ static ResultCode QueryMemory(MemoryInfo* memory_info, PageInfo* page_info, VAdd
 /// Exits the current process
 static void ExitProcess() {
-    LOG_INFO(Kernel_SVC, "Process {} exiting", Core::CurrentProcess()->process_id);
+    auto& current_process = Core::CurrentProcess();
-    ASSERT_MSG(Core::CurrentProcess()->status == ProcessStatus::Running,
+    LOG_INFO(Kernel_SVC, "Process {} exiting", current_process->GetProcessID());
+    ASSERT_MSG(current_process->GetStatus() == ProcessStatus::Running,
               "Process has already exited");
-    Core::CurrentProcess()->status = ProcessStatus::Exited;
+    current_process->PrepareForTermination();
-    auto stop_threads = [](const std::vector<SharedPtr<Thread>>& thread_list) {
-        for (auto& thread : thread_list) {
-            if (thread->owner_process != Core::CurrentProcess())
-                continue;
-            if (thread == GetCurrentThread())
-                continue;
-            // TODO(Subv): When are the other running/ready threads terminated?
-            ASSERT_MSG(thread->status == ThreadStatus::WaitSynchAny ||
-                           thread->status == ThreadStatus::WaitSynchAll,
-                       "Exiting processes with non-waiting threads is currently unimplemented");
-            thread->Stop();
-        }
-    };
-    auto& system = Core::System::GetInstance();
-    stop_threads(system.Scheduler(0)->GetThreadList());
-    stop_threads(system.Scheduler(1)->GetThreadList());
-    stop_threads(system.Scheduler(2)->GetThreadList());
-    stop_threads(system.Scheduler(3)->GetThreadList());
    // Kill the current thread
    GetCurrentThread()->Stop();
@@ -575,14 +596,14 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V
        return ERR_INVALID_THREAD_PRIORITY;
    }
-    SharedPtr<ResourceLimit>& resource_limit = Core::CurrentProcess()->resource_limit;
+    const ResourceLimit& resource_limit = Core::CurrentProcess()->GetResourceLimit();
-    if (resource_limit->GetMaxResourceValue(ResourceType::Priority) > priority) {
+    if (resource_limit.GetMaxResourceValue(ResourceType::Priority) > priority) {
        return ERR_NOT_AUTHORIZED;
    }
    if (processor_id == THREADPROCESSORID_DEFAULT) {
        // Set the target CPU to the one specified in the process' exheader.
-        processor_id = Core::CurrentProcess()->ideal_processor;
+        processor_id = Core::CurrentProcess()->GetDefaultProcessorID();
        ASSERT(processor_id != THREADPROCESSORID_DEFAULT);
    }
@@ -909,10 +930,10 @@ static ResultCode SetThreadCoreMask(Handle thread_handle, u32 core, u64 mask) {
    }
    if (core == static_cast<u32>(THREADPROCESSORID_DEFAULT)) {
-        ASSERT(thread->owner_process->ideal_processor !=
+        ASSERT(thread->owner_process->GetDefaultProcessorID() !=
               static_cast<u8>(THREADPROCESSORID_DEFAULT));
        // Set the target CPU to the one specified in the process' exheader.
-        core = thread->owner_process->ideal_processor;
+        core = thread->owner_process->GetDefaultProcessorID();
        mask = 1ull << core;
    }
@@ -1039,7 +1060,7 @@ static const FunctionDef SVC_Table[] = {
    {0x2B, nullptr, "FlushDataCache"},
    {0x2C, nullptr, "MapPhysicalMemory"},
    {0x2D, nullptr, "UnmapPhysicalMemory"},
-    {0x2E, nullptr, "GetNextThreadInfo"},
+    {0x2E, nullptr, "GetFutureThreadInfo"},
    {0x2F, nullptr, "GetLastThreadInfo"},
    {0x30, nullptr, "GetResourceLimitLimitValue"},
    {0x31, nullptr, "GetResourceLimitCurrentValue"},
@@ -1065,11 +1086,11 @@ static const FunctionDef SVC_Table[] = {
    {0x45, nullptr, "CreateEvent"},
    {0x46, nullptr, "Unknown"},
    {0x47, nullptr, "Unknown"},
-    {0x48, nullptr, "AllocateUnsafeMemory"},
+    {0x48, nullptr, "MapPhysicalMemoryUnsafe"},
-    {0x49, nullptr, "FreeUnsafeMemory"},
+    {0x49, nullptr, "UnmapPhysicalMemoryUnsafe"},
-    {0x4A, nullptr, "SetUnsafeAllocationLimit"},
+    {0x4A, nullptr, "SetUnsafeLimit"},
-    {0x4B, nullptr, "CreateJitMemory"},
+    {0x4B, nullptr, "CreateCodeMemory"},
-    {0x4C, nullptr, "MapJitMemory"},
+    {0x4C, nullptr, "ControlCodeMemory"},
    {0x4D, nullptr, "SleepSystem"},
    {0x4E, nullptr, "ReadWriteRegister"},
    {0x4F, nullptr, "SetProcessActivity"},
@@ -1104,7 +1125,7 @@ static const FunctionDef SVC_Table[] = {
    {0x6C, nullptr, "SetHardwareBreakPoint"},
    {0x6D, nullptr, "GetDebugThreadParam"},
    {0x6E, nullptr, "Unknown"},
-    {0x6F, nullptr, "GetMemoryInfo"},
+    {0x6F, nullptr, "GetSystemInfo"},
    {0x70, nullptr, "CreatePort"},
    {0x71, nullptr, "ManageNamedPort"},
    {0x72, nullptr, "ConnectToPort"},
diff --git a/src/core/hle/kernel/svc_wrap.h b/src/core/hle/kernel/svc_wrap.h
index fea9ba5ea..22712e64f 100644
--- a/src/core/hle/kernel/svc_wrap.h
+++ b/src/core/hle/kernel/svc_wrap.h
@@ -64,6 +64,11 @@ void SvcWrap() {
    FuncReturn(func(Param(0), (s32)Param(1)).raw);
 }
+template <ResultCode func(u64, u32)>
+void SvcWrap() {
+    FuncReturn(func(Param(0), static_cast<u32>(Param(1))).raw);
+}
 template <ResultCode func(u64*, u64)>
 void SvcWrap() {
    u64 param_1 = 0;
diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp
index d4183d6e3..b5c16cfbb 100644
--- a/src/core/hle/kernel/thread.cpp
+++ b/src/core/hle/kernel/thread.cpp
@@ -65,10 +65,7 @@ void Thread::Stop() {
    wait_objects.clear();
    // Mark the TLS slot in the thread's page as free.
-    const u64 tls_page = (tls_address - Memory::TLS_AREA_VADDR) / Memory::PAGE_SIZE;
+    owner_process->FreeTLSSlot(tls_address);
-    const u64 tls_slot =
-        ((tls_address - Memory::TLS_AREA_VADDR) % Memory::PAGE_SIZE) / Memory::TLS_ENTRY_SIZE;
-    Core::CurrentProcess()->tls_slots[tls_page].reset(tls_slot);
 }
 void WaitCurrentThread_Sleep() {
@@ -178,32 +175,6 @@ void Thread::ResumeFromWait() {
 }
 /**
- * Finds a free location for the TLS section of a thread.
- * @param tls_slots The TLS page array of the thread's owner process.
- * Returns a tuple of (page, slot, alloc_needed) where:
- * page: The index of the first allocated TLS page that has free slots.
- * slot: The index of the first free slot in the indicated page.
- * alloc_needed: Whether there's a need to allocate a new TLS page (All pages are full).
- */
-static std::tuple<std::size_t, std::size_t, bool> GetFreeThreadLocalSlot(
-    const std::vector<std::bitset<8>>& tls_slots) {
-    // Iterate over all the allocated pages, and try to find one where not all slots are used.
-    for (std::size_t page = 0; page < tls_slots.size(); ++page) {
-        const auto& page_tls_slots = tls_slots[page];
-        if (!page_tls_slots.all()) {
-            // We found a page with at least one free slot, find which slot it is
-            for (std::size_t slot = 0; slot < page_tls_slots.size(); ++slot) {
-                if (!page_tls_slots.test(slot)) {
-                    return std::make_tuple(page, slot, false);
-                }
-            }
-        }
-    }
-    return std::make_tuple(0, 0, true);
-}
-/**
 * Resets a thread context, making it ready to be scheduled and run by the CPU
 * @param context Thread context to reset
 * @param stack_top Address of the top of the stack
@@ -264,32 +235,7 @@ ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name
    thread->owner_process = owner_process;
    thread->scheduler = Core::System::GetInstance().Scheduler(processor_id);
    thread->scheduler->AddThread(thread, priority);
+    thread->tls_address = thread->owner_process->MarkNextAvailableTLSSlotAsUsed(*thread);
-    // Find the next available TLS index, and mark it as used
-    auto& tls_slots = owner_process->tls_slots;
-    auto [available_page, available_slot, needs_allocation] = GetFreeThreadLocalSlot(tls_slots);
-    if (needs_allocation) {
-        tls_slots.emplace_back(0); // The page is completely available at the start
-        available_page = tls_slots.size() - 1;
-        available_slot = 0; // Use the first slot in the new page
-        // Allocate some memory from the end of the linear heap for this region.
-        const std::size_t offset = thread->tls_memory->size();
-        thread->tls_memory->insert(thread->tls_memory->end(), Memory::PAGE_SIZE, 0);
-        auto& vm_manager = owner_process->vm_manager;
-        vm_manager.RefreshMemoryBlockMappings(thread->tls_memory.get());
-        vm_manager.MapMemoryBlock(Memory::TLS_AREA_VADDR + available_page * Memory::PAGE_SIZE,
-                                  thread->tls_memory, 0, Memory::PAGE_SIZE,
-                                  MemoryState::ThreadLocal);
-    }
-    // Mark the slot as used
-    tls_slots[available_page].set(available_slot);
-    thread->tls_address = Memory::TLS_AREA_VADDR + available_page * Memory::PAGE_SIZE +
-                          available_slot * Memory::TLS_ENTRY_SIZE;
    // TODO(peachum): move to ScheduleThread() when scheduler is added so selected core is used
    // to initialize the context
@@ -311,13 +257,14 @@ void Thread::BoostPriority(u32 priority) {
 }
 SharedPtr<Thread> SetupMainThread(KernelCore& kernel, VAddr entry_point, u32 priority,
-                                  SharedPtr<Process> owner_process) {
+                                  Process& owner_process) {
    // Setup page table so we can write to memory
-    SetCurrentPageTable(&Core::CurrentProcess()->vm_manager.page_table);
+    SetCurrentPageTable(&owner_process.VMManager().page_table);
    // Initialize new "main" thread
+    const VAddr stack_top = owner_process.VMManager().GetTLSIORegionEndAddress();
    auto thread_res = Thread::Create(kernel, "main", entry_point, priority, 0, THREADPROCESSORID_0,
-                                     Memory::STACK_AREA_VADDR_END, std::move(owner_process));
+                                     stack_top, &owner_process);
    SharedPtr<Thread> thread = std::move(thread_res).Unwrap();
diff --git a/src/core/hle/kernel/thread.h b/src/core/hle/kernel/thread.h
index df4748942..4250144c3 100644
--- a/src/core/hle/kernel/thread.h
+++ b/src/core/hle/kernel/thread.h
@@ -62,6 +62,9 @@ enum class ThreadWakeupReason {
 class Thread final : public WaitObject {
 public:
+    using TLSMemory = std::vector<u8>;
+    using TLSMemoryPtr = std::shared_ptr<TLSMemory>;
    /**
     * Creates and returns a new thread. The new thread is immediately scheduled
     * @param kernel The kernel instance this thread will be created under.
@@ -134,6 +137,14 @@ public:
        return thread_id;
    }
+    TLSMemoryPtr& GetTLSMemory() {
+        return tls_memory;
+    }
+    const TLSMemoryPtr& GetTLSMemory() const {
+        return tls_memory;
+    }
    /**
     * Resumes a thread from waiting
     */
@@ -269,7 +280,7 @@ private:
    explicit Thread(KernelCore& kernel);
    ~Thread() override;
-    std::shared_ptr<std::vector<u8>> tls_memory = std::make_shared<std::vector<u8>>();
+    TLSMemoryPtr tls_memory = std::make_shared<TLSMemory>();
 };
 /**
@@ -281,7 +292,7 @@ private:
 * @return A shared pointer to the main thread
 */
 SharedPtr<Thread> SetupMainThread(KernelCore& kernel, VAddr entry_point, u32 priority,
-                                  SharedPtr<Process> owner_process);
+                                  Process& owner_process);
 /**
 * Gets the current thread
diff --git a/src/core/hle/kernel/vm_manager.cpp b/src/core/hle/kernel/vm_manager.cpp
index 608cbd57b..e412309fd 100644
--- a/src/core/hle/kernel/vm_manager.cpp
+++ b/src/core/hle/kernel/vm_manager.cpp
@@ -9,6 +9,7 @@
 #include "common/logging/log.h"
 #include "core/arm/arm_interface.h"
 #include "core/core.h"
+#include "core/file_sys/program_metadata.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/vm_manager.h"
 #include "core/memory.h"
@@ -54,30 +55,32 @@ bool VirtualMemoryArea::CanBeMergedWith(const VirtualMemoryArea& next) const {
 }
 VMManager::VMManager() {
-    Reset();
+    // Default to assuming a 39-bit address space. This way we have a sane
+    // starting point with executables that don't provide metadata.
+    Reset(FileSys::ProgramAddressSpaceType::Is39Bit);
 }
 VMManager::~VMManager() {
-    Reset();
+    Reset(FileSys::ProgramAddressSpaceType::Is39Bit);
 }
-void VMManager::Reset() {
+void VMManager::Reset(FileSys::ProgramAddressSpaceType type) {
-    vma_map.clear();
+    Clear();
+    InitializeMemoryRegionRanges(type);
+    page_table.Resize(address_space_width);
    // Initialize the map with a single free region covering the entire managed space.
    VirtualMemoryArea initial_vma;
-    initial_vma.size = MAX_ADDRESS;
+    initial_vma.size = address_space_end;
    vma_map.emplace(initial_vma.base, initial_vma);
-    page_table.pointers.fill(nullptr);
-    page_table.special_regions.clear();
-    page_table.attributes.fill(Memory::PageType::Unmapped);
    UpdatePageTableForVMA(initial_vma);
 }
 VMManager::VMAHandle VMManager::FindVMA(VAddr target) const {
-    if (target >= MAX_ADDRESS) {
+    if (target >= address_space_end) {
        return vma_map.end();
    } else {
        return std::prev(vma_map.upper_bound(target));
@@ -291,7 +294,7 @@ ResultVal<VMManager::VMAIter> VMManager::CarveVMARange(VAddr target, u64 size) {
    const VAddr target_end = target + size;
    ASSERT(target_end >= target);
-    ASSERT(target_end <= MAX_ADDRESS);
+    ASSERT(target_end <= address_space_end);
    ASSERT(size > 0);
    VMAIter begin_vma = StripIterConstness(FindVMA(target));
@@ -382,6 +385,85 @@ void VMManager::UpdatePageTableForVMA(const VirtualMemoryArea& vma) {
    }
 }
+void VMManager::InitializeMemoryRegionRanges(FileSys::ProgramAddressSpaceType type) {
+    u64 map_region_size = 0;
+    u64 heap_region_size = 0;
+    u64 new_map_region_size = 0;
+    u64 tls_io_region_size = 0;
+    switch (type) {
+    case FileSys::ProgramAddressSpaceType::Is32Bit:
+        address_space_width = 32;
+        code_region_base = 0x200000;
+        code_region_end = code_region_base + 0x3FE00000;
+        map_region_size = 0x40000000;
+        heap_region_size = 0x40000000;
+        break;
+    case FileSys::ProgramAddressSpaceType::Is36Bit:
+        address_space_width = 36;
+        code_region_base = 0x8000000;
+        code_region_end = code_region_base + 0x78000000;
+        map_region_size = 0x180000000;
+        heap_region_size = 0x180000000;
+        break;
+    case FileSys::ProgramAddressSpaceType::Is32BitNoMap:
+        address_space_width = 32;
+        code_region_base = 0x200000;
+        code_region_end = code_region_base + 0x3FE00000;
+        map_region_size = 0;
+        heap_region_size = 0x80000000;
+        break;
+    case FileSys::ProgramAddressSpaceType::Is39Bit:
+        address_space_width = 39;
+        code_region_base = 0x8000000;
+        code_region_end = code_region_base + 0x80000000;
+        map_region_size = 0x1000000000;
+        heap_region_size = 0x180000000;
+        new_map_region_size = 0x80000000;
+        tls_io_region_size = 0x1000000000;
+        break;
+    default:
+        UNREACHABLE_MSG("Invalid address space type specified: {}", static_cast<u32>(type));
+        return;
+    }
+    address_space_base = 0;
+    address_space_end = 1ULL << address_space_width;
+    map_region_base = code_region_end;
+    map_region_end = map_region_base + map_region_size;
+    heap_region_base = map_region_end;
+    heap_region_end = heap_region_base + heap_region_size;
+    new_map_region_base = heap_region_end;
+    new_map_region_end = new_map_region_base + new_map_region_size;
+    tls_io_region_base = new_map_region_end;
+    tls_io_region_end = tls_io_region_base + tls_io_region_size;
+    if (new_map_region_size == 0) {
+        new_map_region_base = address_space_base;
+        new_map_region_end = address_space_end;
+    }
+}
+void VMManager::Clear() {
+    ClearVMAMap();
+    ClearPageTable();
+}
+void VMManager::ClearVMAMap() {
+    vma_map.clear();
+}
+void VMManager::ClearPageTable() {
+    std::fill(page_table.pointers.begin(), page_table.pointers.end(), nullptr);
+    page_table.special_regions.clear();
+    std::fill(page_table.attributes.begin(), page_table.attributes.end(),
+              Memory::PageType::Unmapped);
+}
 u64 VMManager::GetTotalMemoryUsage() const {
    LOG_WARNING(Kernel, "(STUBBED) called");
    return 0xF8000000;
@@ -392,14 +474,80 @@ u64 VMManager::GetTotalHeapUsage() const {
    return 0x0;
 }
-VAddr VMManager::GetAddressSpaceBaseAddr() const {
+VAddr VMManager::GetAddressSpaceBaseAddress() const {
-    LOG_WARNING(Kernel, "(STUBBED) called");
+    return address_space_base;
-    return 0x8000000;
+}
+VAddr VMManager::GetAddressSpaceEndAddress() const {
+    return address_space_end;
 }
 u64 VMManager::GetAddressSpaceSize() const {
-    LOG_WARNING(Kernel, "(STUBBED) called");
+    return address_space_end - address_space_base;
-    return MAX_ADDRESS;
+}
+u64 VMManager::GetAddressSpaceWidth() const {
+    return address_space_width;
+}
+VAddr VMManager::GetCodeRegionBaseAddress() const {
+    return code_region_base;
+}
+VAddr VMManager::GetCodeRegionEndAddress() const {
+    return code_region_end;
+}
+u64 VMManager::GetCodeRegionSize() const {
+    return code_region_end - code_region_base;
+}
+VAddr VMManager::GetHeapRegionBaseAddress() const {
+    return heap_region_base;
+}
+VAddr VMManager::GetHeapRegionEndAddress() const {
+    return heap_region_end;
+}
+u64 VMManager::GetHeapRegionSize() const {
+    return heap_region_end - heap_region_base;
+}
+VAddr VMManager::GetMapRegionBaseAddress() const {
+    return map_region_base;
+}
+VAddr VMManager::GetMapRegionEndAddress() const {
+    return map_region_end;
+}
+u64 VMManager::GetMapRegionSize() const {
+    return map_region_end - map_region_base;
+}
+VAddr VMManager::GetNewMapRegionBaseAddress() const {
+    return new_map_region_base;
+}
+VAddr VMManager::GetNewMapRegionEndAddress() const {
+    return new_map_region_end;
+}
+u64 VMManager::GetNewMapRegionSize() const {
+    return new_map_region_end - new_map_region_base;
+}
+VAddr VMManager::GetTLSIORegionBaseAddress() const {
+    return tls_io_region_base;
+}
+VAddr VMManager::GetTLSIORegionEndAddress() const {
+    return tls_io_region_end;
+}
+u64 VMManager::GetTLSIORegionSize() const {
+    return tls_io_region_end - tls_io_region_base;
 }
 } // namespace Kernel
diff --git a/src/core/hle/kernel/vm_manager.h b/src/core/hle/kernel/vm_manager.h
index de75036c0..015559a64 100644
--- a/src/core/hle/kernel/vm_manager.h
+++ b/src/core/hle/kernel/vm_manager.h
@@ -12,6 +12,10 @@
 #include "core/memory.h"
 #include "core/memory_hook.h"
+namespace FileSys {
+enum class ProgramAddressSpaceType : u8;
+}
 namespace Kernel {
 enum class VMAType : u8 {
@@ -111,12 +115,6 @@ struct VirtualMemoryArea {
 class VMManager final {
 public:
    /**
-     * The maximum amount of address space managed by the kernel.
-     * @todo This was selected arbitrarily, and should be verified for Switch OS.
-     */
-    static constexpr VAddr MAX_ADDRESS{0x1000000000ULL};
-    /**
     * A map covering the entirety of the managed address space, keyed by the `base` field of each
     * VMA. It must always be modified by splitting or merging VMAs, so that the invariant
     * `elem.base + elem.size == next.base` is preserved, and mergeable regions must always be
@@ -130,7 +128,7 @@ public:
    ~VMManager();
    /// Clears the address space map, re-initializing with a single free area.
-    void Reset();
+    void Reset(FileSys::ProgramAddressSpaceType type);
    /// Finds the VMA in which the given address is included in, or `vma_map.end()`.
    VMAHandle FindVMA(VAddr target) const;
@@ -195,12 +193,63 @@ public:
    /// Gets the total heap usage, used by svcGetInfo
    u64 GetTotalHeapUsage() const;
-    /// Gets the total address space base address, used by svcGetInfo
+    /// Gets the address space base address
-    VAddr GetAddressSpaceBaseAddr() const;
+    VAddr GetAddressSpaceBaseAddress() const;
-    /// Gets the total address space address size, used by svcGetInfo
+    /// Gets the address space end address
+    VAddr GetAddressSpaceEndAddress() const;
+    /// Gets the total address space address size in bytes
    u64 GetAddressSpaceSize() const;
+    /// Gets the address space width in bits.
+    u64 GetAddressSpaceWidth() const;
+    /// Gets the base address of the code region.
+    VAddr GetCodeRegionBaseAddress() const;
+    /// Gets the end address of the code region.
+    VAddr GetCodeRegionEndAddress() const;
+    /// Gets the total size of the code region in bytes.
+    u64 GetCodeRegionSize() const;
+    /// Gets the base address of the heap region.
+    VAddr GetHeapRegionBaseAddress() const;
+    /// Gets the end address of the heap region;
+    VAddr GetHeapRegionEndAddress() const;
+    /// Gets the total size of the heap region in bytes.
+    u64 GetHeapRegionSize() const;
+    /// Gets the base address of the map region.
+    VAddr GetMapRegionBaseAddress() const;
+    /// Gets the end address of the map region.
+    VAddr GetMapRegionEndAddress() const;
+    /// Gets the total size of the map region in bytes.
+    u64 GetMapRegionSize() const;
+    /// Gets the base address of the new map region.
+    VAddr GetNewMapRegionBaseAddress() const;
+    /// Gets the end address of the new map region.
+    VAddr GetNewMapRegionEndAddress() const;
+    /// Gets the total size of the new map region in bytes.
+    u64 GetNewMapRegionSize() const;
+    /// Gets the base address of the TLS IO region.
+    VAddr GetTLSIORegionBaseAddress() const;
+    /// Gets the end address of the TLS IO region.
+    VAddr GetTLSIORegionEndAddress() const;
+    /// Gets the total size of the TLS IO region in bytes.
+    u64 GetTLSIORegionSize() const;
    /// Each VMManager has its own page table, which is set as the main one when the owning process
    /// is scheduled.
    Memory::PageTable page_table;
@@ -240,5 +289,36 @@ private:
    /// Updates the pages corresponding to this VMA so they match the VMA's attributes.
    void UpdatePageTableForVMA(const VirtualMemoryArea& vma);
+    /// Initializes memory region ranges to adhere to a given address space type.
+    void InitializeMemoryRegionRanges(FileSys::ProgramAddressSpaceType type);
+    /// Clears the underlying map and page table.
+    void Clear();
+    /// Clears out the VMA map, unmapping any previously mapped ranges.
+    void ClearVMAMap();
+    /// Clears out the page table
+    void ClearPageTable();
+    u32 address_space_width = 0;
+    VAddr address_space_base = 0;
+    VAddr address_space_end = 0;
+    VAddr code_region_base = 0;
+    VAddr code_region_end = 0;
+    VAddr heap_region_base = 0;
+    VAddr heap_region_end = 0;
+    VAddr map_region_base = 0;
+    VAddr map_region_end = 0;
+    VAddr new_map_region_base = 0;
+    VAddr new_map_region_end = 0;
+    VAddr tls_io_region_base = 0;
+    VAddr tls_io_region_end = 0;
 };
 } // namespace Kernel
diff --git a/src/core/hle/kernel/wait_object.h b/src/core/hle/kernel/wait_object.h
index 0bd97133c..f4367ee28 100644
--- a/src/core/hle/kernel/wait_object.h
+++ b/src/core/hle/kernel/wait_object.h
@@ -69,7 +69,7 @@ private:
 template <>
 inline SharedPtr<WaitObject> DynamicObjectCast<WaitObject>(SharedPtr<Object> object) {
    if (object != nullptr && object->IsWaitable()) {
-        return boost::static_pointer_cast<WaitObject>(std::move(object));
+        return boost::static_pointer_cast<WaitObject>(object);
    }
    return nullptr;
 }
diff --git a/src/core/hle/service/acc/acc.cpp b/src/core/hle/service/acc/acc.cpp
index 4d4eb542e..e61748ca3 100644
--- a/src/core/hle/service/acc/acc.cpp
+++ b/src/core/hle/service/acc/acc.cpp
@@ -130,11 +130,10 @@ private:
    void GetAccountId(Kernel::HLERequestContext& ctx) {
        LOG_WARNING(Service_ACC, "(STUBBED) called");
-        // TODO(Subv): Find out what this actually does and implement it. Stub it as an error for
+        // Should return a nintendo account ID
-        // now since we do not implement NNID. Returning a bogus id here will cause games to send
+        IPC::ResponseBuilder rb{ctx, 4};
-        // invalid IPC requests after ListOpenUsers is called.
+        rb.Push(RESULT_SUCCESS);
-        IPC::ResponseBuilder rb{ctx, 2};
+        rb.PushRaw<u64>(1);
-        rb.Push(ResultCode(-1));
    }
 };
diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp
index 9c975325a..69bfce1c1 100644
--- a/src/core/hle/service/am/am.cpp
+++ b/src/core/hle/service/am/am.cpp
@@ -462,7 +462,7 @@ private:
        std::memcpy(&buffer[offset], data.data(), data.size());
-        IPC::ResponseBuilder rb{rp.MakeBuilder(2, 0, 0)};
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
        LOG_DEBUG(Service_AM, "called, offset={}", offset);
@@ -478,7 +478,7 @@ private:
        ctx.WriteBuffer(buffer.data() + offset, size);
-        IPC::ResponseBuilder rb{rp.MakeBuilder(2, 0, 0)};
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
        LOG_DEBUG(Service_AM, "called, offset={}", offset);
@@ -568,7 +568,7 @@ private:
        IPC::RequestParser rp{ctx};
        storage_stack.push(rp.PopIpcInterface<AM::IStorage>());
-        IPC::ResponseBuilder rb{rp.MakeBuilder(2, 0, 0)};
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
        LOG_DEBUG(Service_AM, "called");
@@ -616,7 +616,7 @@ void ILibraryAppletCreator::CreateStorage(Kernel::HLERequestContext& ctx) {
    const u64 size{rp.Pop<u64>()};
    std::vector<u8> buffer(size);
-    IPC::ResponseBuilder rb{rp.MakeBuilder(2, 0, 1)};
+    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
    rb.Push(RESULT_SUCCESS);
    rb.PushIpcInterface<AM::IStorage>(std::move(buffer));
diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp
index 80a002322..ff1edefbb 100644
--- a/src/core/hle/service/audio/audout_u.cpp
+++ b/src/core/hle/service/audio/audout_u.cpp
@@ -190,7 +190,7 @@ void AudOutU::ListAudioOutsImpl(Kernel::HLERequestContext& ctx) {
    ctx.WriteBuffer(DefaultDevice);
-    IPC::ResponseBuilder rb = rp.MakeBuilder(3, 0, 0);
+    IPC::ResponseBuilder rb{ctx, 3};
    rb.Push(RESULT_SUCCESS);
    rb.Push<u32>(1); // Amount of audio devices
diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp
index e84c4fa2b..6073f4ecd 100644
--- a/src/core/hle/service/audio/audren_u.cpp
+++ b/src/core/hle/service/audio/audren_u.cpp
@@ -10,6 +10,7 @@
 #include "common/alignment.h"
 #include "common/common_funcs.h"
 #include "common/logging/log.h"
+#include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/kernel/hle_ipc.h"
@@ -25,7 +26,7 @@ public:
            {0, &IAudioRenderer::GetAudioRendererSampleRate, "GetAudioRendererSampleRate"},
            {1, &IAudioRenderer::GetAudioRendererSampleCount, "GetAudioRendererSampleCount"},
            {2, &IAudioRenderer::GetAudioRendererMixBufferCount, "GetAudioRendererMixBufferCount"},
-            {3, nullptr, "GetAudioRendererState"},
+            {3, &IAudioRenderer::GetAudioRendererState, "GetAudioRendererState"},
            {4, &IAudioRenderer::RequestUpdateAudioRenderer, "RequestUpdateAudioRenderer"},
            {5, &IAudioRenderer::StartAudioRenderer, "StartAudioRenderer"},
            {6, &IAudioRenderer::StopAudioRenderer, "StopAudioRenderer"},
@@ -62,6 +63,13 @@ private:
        LOG_DEBUG(Service_Audio, "called");
    }
+    void GetAudioRendererState(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(static_cast<u32>(renderer->GetStreamState()));
+        LOG_DEBUG(Service_Audio, "called");
+    }
    void GetAudioRendererMixBufferCount(Kernel::HLERequestContext& ctx) {
        IPC::ResponseBuilder rb{ctx, 3};
        rb.Push(RESULT_SUCCESS);
@@ -137,7 +145,7 @@ private:
        constexpr std::array<char, 15> audio_interface{{"AudioInterface"}};
        ctx.WriteBuffer(audio_interface);
-        IPC::ResponseBuilder rb = rp.MakeBuilder(3, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 3};
        rb.Push(RESULT_SUCCESS);
        rb.Push<u32>(1);
    }
@@ -151,7 +159,7 @@ private:
        auto file_buffer = ctx.ReadBuffer();
        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
    }
@@ -162,7 +170,7 @@ private:
        constexpr std::array<char, 12> audio_interface{{"AudioDevice"}};
        ctx.WriteBuffer(audio_interface);
-        IPC::ResponseBuilder rb = rp.MakeBuilder(3, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 3};
        rb.Push(RESULT_SUCCESS);
        rb.Push<u32>(1);
    }
diff --git a/src/core/hle/service/fatal/fatal.cpp b/src/core/hle/service/fatal/fatal.cpp
index b436ce4e6..2f15ac2a6 100644
--- a/src/core/hle/service/fatal/fatal.cpp
+++ b/src/core/hle/service/fatal/fatal.cpp
@@ -2,8 +2,17 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+#include <array>
+#include <cstring>
+#include <ctime>
+#include <fmt/time.h>
+#include "common/file_util.h"
 #include "common/logging/log.h"
+#include "common/scm_rev.h"
+#include "common/swap.h"
+#include "core/core.h"
 #include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/process.h"
 #include "core/hle/service/fatal/fatal.h"
 #include "core/hle/service/fatal/fatal_p.h"
 #include "core/hle/service/fatal/fatal_u.h"
@@ -15,16 +24,142 @@ Module::Interface::Interface(std::shared_ptr<Module> module, const char* name)
 Module::Interface::~Interface() = default;
+struct FatalInfo {
+    std::array<u64_le, 31> registers{}; // TODO(ogniK): See if this actually is registers or
+                                        // not(find a game which has non zero valeus)
+    u64_le unk0{};
+    u64_le unk1{};
+    u64_le unk2{};
+    u64_le unk3{};
+    u64_le unk4{};
+    u64_le unk5{};
+    u64_le unk6{};
+    std::array<u64_le, 32> backtrace{};
+    u64_le unk7{};
+    u64_le unk8{};
+    u32_le backtrace_size{};
+    u32_le unk9{};
+    u32_le unk10{}; // TODO(ogniK): Is this even used or is it just padding?
+};
+static_assert(sizeof(FatalInfo) == 0x250, "FatalInfo is an invalid size");
+enum class FatalType : u32 {
+    ErrorReportAndScreen = 0,
+    ErrorReport = 1,
+    ErrorScreen = 2,
+};
+static void GenerateErrorReport(ResultCode error_code, const FatalInfo& info) {
+    const auto title_id = Core::CurrentProcess()->GetTitleID();
+    std::string crash_report =
+        fmt::format("Yuzu {}-{} crash report\n"
+                    "Title ID:                        {:016x}\n"
+                    "Result:                          0x{:X} ({:04}-{:04d})\n"
+                    "\n",
+                    Common::g_scm_branch, Common::g_scm_desc, title_id, error_code.raw,
+                    2000 + static_cast<u32>(error_code.module.Value()),
+                    static_cast<u32>(error_code.description.Value()), info.unk8, info.unk7);
+    if (info.backtrace_size != 0x0) {
+        crash_report += "Registers:\n";
+        // TODO(ogniK): This is just a guess, find a game which actually has non zero values
+        for (size_t i = 0; i < info.registers.size(); i++) {
+            crash_report +=
+                fmt::format("    X[{:02d}]:                       {:016x}\n", i, info.registers[i]);
+        }
+        crash_report += fmt::format("    Unknown 0:                   {:016x}\n", info.unk0);
+        crash_report += fmt::format("    Unknown 1:                   {:016x}\n", info.unk1);
+        crash_report += fmt::format("    Unknown 2:                   {:016x}\n", info.unk2);
+        crash_report += fmt::format("    Unknown 3:                   {:016x}\n", info.unk3);
+        crash_report += fmt::format("    Unknown 4:                   {:016x}\n", info.unk4);
+        crash_report += fmt::format("    Unknown 5:                   {:016x}\n", info.unk5);
+        crash_report += fmt::format("    Unknown 6:                   {:016x}\n", info.unk6);
+        crash_report += "\nBacktrace:\n";
+        for (size_t i = 0; i < info.backtrace_size; i++) {
+            crash_report +=
+                fmt::format("    Backtrace[{:02d}]:               {:016x}\n", i, info.backtrace[i]);
+        }
+        crash_report += fmt::format("\nUnknown 7:                       0x{:016x}\n", info.unk7);
+        crash_report += fmt::format("Unknown 8:                       0x{:016x}\n", info.unk8);
+        crash_report += fmt::format("Unknown 9:                       0x{:016x}\n", info.unk9);
+        crash_report += fmt::format("Unknown 10:                      0x{:016x}\n", info.unk10);
+    }
+    LOG_ERROR(Service_Fatal, "{}", crash_report);
+    const std::string crashreport_dir =
+        FileUtil::GetUserPath(FileUtil::UserPath::LogDir) + "crash_logs";
+    if (!FileUtil::CreateFullPath(crashreport_dir)) {
+        LOG_ERROR(
+            Service_Fatal,
+            "Unable to create crash report directory. Possible log directory permissions issue.");
+        return;
+    }
+    const std::time_t t = std::time(nullptr);
+    const std::string crashreport_filename =
+        fmt::format("{}/{:016x}-{:%F-%H%M%S}.log", crashreport_dir, title_id, *std::localtime(&t));
+    auto file = FileUtil::IOFile(crashreport_filename, "wb");
+    if (file.IsOpen()) {
+        file.WriteString(crash_report);
+        LOG_ERROR(Service_Fatal, "Saving error report to {}", crashreport_filename);
+    } else {
+        LOG_ERROR(Service_Fatal, "Failed to save error report to {}", crashreport_filename);
+    }
+}
+static void ThrowFatalError(ResultCode error_code, FatalType fatal_type, const FatalInfo& info) {
+    LOG_ERROR(Service_Fatal, "Threw fatal error type {}", static_cast<u32>(fatal_type));
+    switch (fatal_type) {
+    case FatalType::ErrorReportAndScreen:
+        GenerateErrorReport(error_code, info);
+        [[fallthrough]];
+    case FatalType::ErrorScreen:
+        // Since we have no fatal:u error screen. We should just kill execution instead
+        ASSERT(false);
+        break;
+        // Should not throw a fatal screen but should generate an error report
+    case FatalType::ErrorReport:
+        GenerateErrorReport(error_code, info);
+        break;
+    };
+}
+void Module::Interface::ThrowFatal(Kernel::HLERequestContext& ctx) {
+    LOG_ERROR(Service_Fatal, "called");
+    IPC::RequestParser rp{ctx};
+    auto error_code = rp.Pop<ResultCode>();
+    ThrowFatalError(error_code, FatalType::ErrorScreen, {});
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
 void Module::Interface::ThrowFatalWithPolicy(Kernel::HLERequestContext& ctx) {
+    LOG_ERROR(Service_Fatal, "called");
    IPC::RequestParser rp(ctx);
-    u32 error_code = rp.Pop<u32>();
+    auto error_code = rp.Pop<ResultCode>();
-    LOG_WARNING(Service_Fatal, "(STUBBED) called, error_code=0x{:X}", error_code);
+    auto fatal_type = rp.PopEnum<FatalType>();
+    ThrowFatalError(error_code, fatal_type, {}); // No info is passed with ThrowFatalWithPolicy
    IPC::ResponseBuilder rb{ctx, 2};
    rb.Push(RESULT_SUCCESS);
 }
 void Module::Interface::ThrowFatalWithCpuContext(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_Fatal, "(STUBBED) called");
+    LOG_ERROR(Service_Fatal, "called");
+    IPC::RequestParser rp(ctx);
+    auto error_code = rp.Pop<ResultCode>();
+    auto fatal_type = rp.PopEnum<FatalType>();
+    auto fatal_info = ctx.ReadBuffer();
+    FatalInfo info{};
+    ASSERT_MSG(fatal_info.size() == sizeof(FatalInfo), "Invalid fatal info buffer size!");
+    std::memcpy(&info, fatal_info.data(), sizeof(FatalInfo));
+    ThrowFatalError(error_code, fatal_type, info);
    IPC::ResponseBuilder rb{ctx, 2};
    rb.Push(RESULT_SUCCESS);
 }
diff --git a/src/core/hle/service/fatal/fatal.h b/src/core/hle/service/fatal/fatal.h
index 4d9a5be52..09371ff7f 100644
--- a/src/core/hle/service/fatal/fatal.h
+++ b/src/core/hle/service/fatal/fatal.h
@@ -15,6 +15,7 @@ public:
        explicit Interface(std::shared_ptr<Module> module, const char* name);
        ~Interface() override;
+        void ThrowFatal(Kernel::HLERequestContext& ctx);
        void ThrowFatalWithPolicy(Kernel::HLERequestContext& ctx);
        void ThrowFatalWithCpuContext(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/fatal/fatal_u.cpp b/src/core/hle/service/fatal/fatal_u.cpp
index befc307cf..1572a2051 100644
--- a/src/core/hle/service/fatal/fatal_u.cpp
+++ b/src/core/hle/service/fatal/fatal_u.cpp
@@ -8,7 +8,7 @@ namespace Service::Fatal {
 Fatal_U::Fatal_U(std::shared_ptr<Module> module) : Module::Interface(std::move(module), "fatal:u") {
    static const FunctionInfo functions[] = {
-        {0, nullptr, "ThrowFatal"},
+        {0, &Fatal_U::ThrowFatal, "ThrowFatal"},
        {1, &Fatal_U::ThrowFatalWithPolicy, "ThrowFatalWithPolicy"},
        {2, &Fatal_U::ThrowFatalWithCpuContext, "ThrowFatalWithCpuContext"},
    };
diff --git a/src/core/hle/service/filesystem/filesystem.cpp b/src/core/hle/service/filesystem/filesystem.cpp
index 5c4971724..aed2abb71 100644
--- a/src/core/hle/service/filesystem/filesystem.cpp
+++ b/src/core/hle/service/filesystem/filesystem.cpp
@@ -197,7 +197,7 @@ ResultVal<FileSys::VirtualDir> VfsDirectoryServiceWrapper::OpenDirectory(const s
    auto dir = GetDirectoryRelativeWrapped(backing, path);
    if (dir == nullptr) {
        // TODO(DarkLordZach): Find a better error code for this
-        return ResultCode(-1);
+        return FileSys::ERROR_PATH_NOT_FOUND;
    }
    return MakeResult(dir);
 }
@@ -343,6 +343,15 @@ std::shared_ptr<FileSys::RegisteredCache> GetSDMCContents() {
    return sdmc_factory->GetSDMCContents();
 }
+FileSys::VirtualDir GetModificationLoadRoot(u64 title_id) {
+    LOG_TRACE(Service_FS, "Opening mod load root for tid={:016X}", title_id);
+    if (bis_factory == nullptr)
+        return nullptr;
+    return bis_factory->GetModificationLoadRoot(title_id);
+}
 void CreateFactories(const FileSys::VirtualFilesystem& vfs, bool overwrite) {
    if (overwrite) {
        bis_factory = nullptr;
@@ -354,9 +363,11 @@ void CreateFactories(const FileSys::VirtualFilesystem& vfs, bool overwrite) {
                                             FileSys::Mode::ReadWrite);
    auto sd_directory = vfs->OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir),
                                           FileSys::Mode::ReadWrite);
+    auto load_directory = vfs->OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir),
+                                             FileSys::Mode::ReadWrite);
    if (bis_factory == nullptr)
-        bis_factory = std::make_unique<FileSys::BISFactory>(nand_directory);
+        bis_factory = std::make_unique<FileSys::BISFactory>(nand_directory, load_directory);
    if (save_data_factory == nullptr)
        save_data_factory = std::make_unique<FileSys::SaveDataFactory>(std::move(nand_directory));
    if (sdmc_factory == nullptr)
diff --git a/src/core/hle/service/filesystem/filesystem.h b/src/core/hle/service/filesystem/filesystem.h
index aab65a2b8..7039a2247 100644
--- a/src/core/hle/service/filesystem/filesystem.h
+++ b/src/core/hle/service/filesystem/filesystem.h
@@ -52,6 +52,8 @@ std::shared_ptr<FileSys::RegisteredCache> GetSystemNANDContents();
 std::shared_ptr<FileSys::RegisteredCache> GetUserNANDContents();
 std::shared_ptr<FileSys::RegisteredCache> GetSDMCContents();
+FileSys::VirtualDir GetModificationLoadRoot(u64 title_id);
 // Creates the SaveData, SDMC, and BIS Factories. Should be called once and before any function
 // above is called.
 void CreateFactories(const FileSys::VirtualFilesystem& vfs, bool overwrite = true);
diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp
index 256c49bfc..7c6b0a4e6 100644
--- a/src/core/hle/service/hid/hid.cpp
+++ b/src/core/hle/service/hid/hid.cpp
@@ -338,7 +338,7 @@ public:
            {106, &Hid::AcquireNpadStyleSetUpdateEventHandle, "AcquireNpadStyleSetUpdateEventHandle"},
            {107, &Hid::DisconnectNpad, "DisconnectNpad"},
            {108, &Hid::GetPlayerLedPattern, "GetPlayerLedPattern"},
-            {109, nullptr, "ActivateNpadWithRevision"},
+            {109, &Hid::ActivateNpadWithRevision, "ActivateNpadWithRevision"},
            {120, &Hid::SetNpadJoyHoldType, "SetNpadJoyHoldType"},
            {121, &Hid::GetNpadJoyHoldType, "GetNpadJoyHoldType"},
            {122, &Hid::SetNpadJoyAssignmentModeSingleByDefault, "SetNpadJoyAssignmentModeSingleByDefault"},
@@ -603,6 +603,12 @@ private:
        rb.Push(RESULT_SUCCESS);
        LOG_WARNING(Service_HID, "(STUBBED) called");
    }
+    void ActivateNpadWithRevision(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_WARNING(Service_HID, "(STUBBED) called");
+    }
 };
 class HidDbg final : public ServiceFramework<HidDbg> {
diff --git a/src/core/hle/service/hid/irs.cpp b/src/core/hle/service/hid/irs.cpp
index e587ad0d8..872e3c344 100644
--- a/src/core/hle/service/hid/irs.cpp
+++ b/src/core/hle/service/hid/irs.cpp
@@ -2,6 +2,11 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+#include "common/swap.h"
+#include "core/core.h"
+#include "core/core_timing.h"
+#include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/shared_memory.h"
 #include "core/hle/service/hid/irs.h"
 namespace Service::HID {
@@ -9,28 +14,145 @@ namespace Service::HID {
 IRS::IRS() : ServiceFramework{"irs"} {
    // clang-format off
    static const FunctionInfo functions[] = {
-        {302, nullptr, "ActivateIrsensor"},
+        {302, &IRS::ActivateIrsensor, "ActivateIrsensor"},
-        {303, nullptr, "DeactivateIrsensor"},
+        {303, &IRS::DeactivateIrsensor, "DeactivateIrsensor"},
-        {304, nullptr, "GetIrsensorSharedMemoryHandle"},
+        {304, &IRS::GetIrsensorSharedMemoryHandle, "GetIrsensorSharedMemoryHandle"},
-        {305, nullptr, "StopImageProcessor"},
+        {305, &IRS::StopImageProcessor, "StopImageProcessor"},
-        {306, nullptr, "RunMomentProcessor"},
+        {306, &IRS::RunMomentProcessor, "RunMomentProcessor"},
-        {307, nullptr, "RunClusteringProcessor"},
+        {307, &IRS::RunClusteringProcessor, "RunClusteringProcessor"},
-        {308, nullptr, "RunImageTransferProcessor"},
+        {308, &IRS::RunImageTransferProcessor, "RunImageTransferProcessor"},
-        {309, nullptr, "GetImageTransferProcessorState"},
+        {309, &IRS::GetImageTransferProcessorState, "GetImageTransferProcessorState"},
-        {310, nullptr, "RunTeraPluginProcessor"},
+        {310, &IRS::RunTeraPluginProcessor, "RunTeraPluginProcessor"},
-        {311, nullptr, "GetNpadIrCameraHandle"},
+        {311, &IRS::GetNpadIrCameraHandle, "GetNpadIrCameraHandle"},
-        {312, nullptr, "RunPointingProcessor"},
+        {312, &IRS::RunPointingProcessor, "RunPointingProcessor"},
-        {313, nullptr, "SuspendImageProcessor"},
+        {313, &IRS::SuspendImageProcessor, "SuspendImageProcessor"},
-        {314, nullptr, "CheckFirmwareVersion"},
+        {314, &IRS::CheckFirmwareVersion, "CheckFirmwareVersion"},
-        {315, nullptr, "SetFunctionLevel"},
+        {315, &IRS::SetFunctionLevel, "SetFunctionLevel"},
-        {316, nullptr, "RunImageTransferExProcessor"},
+        {316, &IRS::RunImageTransferExProcessor, "RunImageTransferExProcessor"},
-        {317, nullptr, "RunIrLedProcessor"},
+        {317, &IRS::RunIrLedProcessor, "RunIrLedProcessor"},
-        {318, nullptr, "StopImageProcessorAsync"},
+        {318, &IRS::StopImageProcessorAsync, "StopImageProcessorAsync"},
-        {319, nullptr, "ActivateIrsensorWithFunctionLevel"},
+        {319, &IRS::ActivateIrsensorWithFunctionLevel, "ActivateIrsensorWithFunctionLevel"},
    };
    // clang-format on
    RegisterHandlers(functions);
+    auto& kernel = Core::System::GetInstance().Kernel();
+    shared_mem = Kernel::SharedMemory::Create(
+        kernel, nullptr, 0x8000, Kernel::MemoryPermission::ReadWrite,
+        Kernel::MemoryPermission::Read, 0, Kernel::MemoryRegion::BASE, "IRS:SharedMemory");
+}
+void IRS::ActivateIrsensor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::DeactivateIrsensor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::GetIrsensorSharedMemoryHandle(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushCopyObjects(shared_mem);
+    LOG_DEBUG(Service_IRS, "called");
+}
+void IRS::StopImageProcessor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::RunMomentProcessor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::RunClusteringProcessor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::RunImageTransferProcessor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::GetImageTransferProcessorState(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 5};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushRaw<u64>(CoreTiming::GetTicks());
+    rb.PushRaw<u32>(0);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::RunTeraPluginProcessor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::GetNpadIrCameraHandle(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushRaw<u32>(device_handle);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::RunPointingProcessor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::SuspendImageProcessor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::CheckFirmwareVersion(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::SetFunctionLevel(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::RunImageTransferExProcessor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::RunIrLedProcessor(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::StopImageProcessorAsync(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
+}
+void IRS::ActivateIrsensorWithFunctionLevel(Kernel::HLERequestContext& ctx) {
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+    LOG_WARNING(Service_IRS, "(STUBBED) called");
 }
 IRS::~IRS() = default;
diff --git a/src/core/hle/service/hid/irs.h b/src/core/hle/service/hid/irs.h
index 6fb16b45d..12de6bfb3 100644
--- a/src/core/hle/service/hid/irs.h
+++ b/src/core/hle/service/hid/irs.h
@@ -4,14 +4,41 @@
 #pragma once
+#include "core/hle/kernel/object.h"
 #include "core/hle/service/service.h"
+namespace Kernel {
+class SharedMemory;
+}
 namespace Service::HID {
 class IRS final : public ServiceFramework<IRS> {
 public:
    explicit IRS();
    ~IRS() override;
+private:
+    void ActivateIrsensor(Kernel::HLERequestContext& ctx);
+    void DeactivateIrsensor(Kernel::HLERequestContext& ctx);
+    void GetIrsensorSharedMemoryHandle(Kernel::HLERequestContext& ctx);
+    void StopImageProcessor(Kernel::HLERequestContext& ctx);
+    void RunMomentProcessor(Kernel::HLERequestContext& ctx);
+    void RunClusteringProcessor(Kernel::HLERequestContext& ctx);
+    void RunImageTransferProcessor(Kernel::HLERequestContext& ctx);
+    void GetImageTransferProcessorState(Kernel::HLERequestContext& ctx);
+    void RunTeraPluginProcessor(Kernel::HLERequestContext& ctx);
+    void GetNpadIrCameraHandle(Kernel::HLERequestContext& ctx);
+    void RunPointingProcessor(Kernel::HLERequestContext& ctx);
+    void SuspendImageProcessor(Kernel::HLERequestContext& ctx);
+    void CheckFirmwareVersion(Kernel::HLERequestContext& ctx);
+    void SetFunctionLevel(Kernel::HLERequestContext& ctx);
+    void RunImageTransferExProcessor(Kernel::HLERequestContext& ctx);
+    void RunIrLedProcessor(Kernel::HLERequestContext& ctx);
+    void StopImageProcessorAsync(Kernel::HLERequestContext& ctx);
+    void ActivateIrsensorWithFunctionLevel(Kernel::HLERequestContext& ctx);
+    Kernel::SharedPtr<Kernel::SharedMemory> shared_mem;
+    const u32 device_handle{0xABCD};
 };
 class IRS_SYS final : public ServiceFramework<IRS_SYS> {
diff --git a/src/core/hle/service/nfp/nfp.cpp b/src/core/hle/service/nfp/nfp.cpp
index f8d2127d9..8c07a05c2 100644
--- a/src/core/hle/service/nfp/nfp.cpp
+++ b/src/core/hle/service/nfp/nfp.cpp
@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 #include "common/logging/log.h"
+#include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/event.h"
 #include "core/hle/service/hid/hid.h"
diff --git a/src/core/hle/service/nifm/nifm.cpp b/src/core/hle/service/nifm/nifm.cpp
index ed4f5f539..10611ed6a 100644
--- a/src/core/hle/service/nifm/nifm.cpp
+++ b/src/core/hle/service/nifm/nifm.cpp
@@ -31,7 +31,7 @@ public:
            {1, &IRequest::GetResult, "GetResult"},
            {2, &IRequest::GetSystemEventReadableHandles, "GetSystemEventReadableHandles"},
            {3, &IRequest::Cancel, "Cancel"},
-            {4, nullptr, "Submit"},
+            {4, &IRequest::Submit, "Submit"},
            {5, nullptr, "SetRequirement"},
            {6, nullptr, "SetRequirementPreset"},
            {8, nullptr, "SetPriority"},
@@ -61,6 +61,12 @@ public:
    }
 private:
+    void Submit(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NIFM, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
    void GetRequestState(Kernel::HLERequestContext& ctx) {
        LOG_WARNING(Service_NIFM, "(STUBBED) called");
        IPC::ResponseBuilder rb{ctx, 3};
@@ -114,10 +120,11 @@ public:
 private:
    void GetClientId(Kernel::HLERequestContext& ctx) {
+        static constexpr u32 client_id = 1;
        LOG_WARNING(Service_NIFM, "(STUBBED) called");
        IPC::ResponseBuilder rb{ctx, 4};
        rb.Push(RESULT_SUCCESS);
-        rb.Push<u64>(0);
+        rb.Push<u64>(client_id); // Client ID needs to be non zero otherwise it's considered invalid
    }
    void CreateScanRequest(Kernel::HLERequestContext& ctx) {
        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
@@ -141,10 +148,16 @@ private:
        rb.Push(RESULT_SUCCESS);
    }
    void CreateTemporaryNetworkProfile(Kernel::HLERequestContext& ctx) {
-        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        ASSERT_MSG(ctx.GetReadBufferSize() == 0x17c, "NetworkProfileData is not the correct size");
+        u128 uuid{};
+        auto buffer = ctx.ReadBuffer();
+        std::memcpy(&uuid, buffer.data() + 8, sizeof(u128));
+        IPC::ResponseBuilder rb{ctx, 6, 0, 1};
        rb.Push(RESULT_SUCCESS);
        rb.PushIpcInterface<INetworkProfile>();
+        rb.PushRaw<u128>(uuid);
        LOG_DEBUG(Service_NIFM, "called");
    }
diff --git a/src/core/hle/service/nim/nim.cpp b/src/core/hle/service/nim/nim.cpp
index bd05b0a70..261ad539c 100644
--- a/src/core/hle/service/nim/nim.cpp
+++ b/src/core/hle/service/nim/nim.cpp
@@ -2,6 +2,11 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+#include <chrono>
+#include <ctime>
+#include "core/core.h"
+#include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/event.h"
 #include "core/hle/service/nim/nim.h"
 #include "core/hle/service/service.h"
 #include "core/hle/service/sm/sm.h"
@@ -100,19 +105,111 @@ public:
    }
 };
+class IEnsureNetworkClockAvailabilityService final
+    : public ServiceFramework<IEnsureNetworkClockAvailabilityService> {
+public:
+    IEnsureNetworkClockAvailabilityService()
+        : ServiceFramework("IEnsureNetworkClockAvailabilityService") {
+        static const FunctionInfo functions[] = {
+            {0, &IEnsureNetworkClockAvailabilityService::StartTask, "StartTask"},
+            {1, &IEnsureNetworkClockAvailabilityService::GetFinishNotificationEvent,
+             "GetFinishNotificationEvent"},
+            {2, &IEnsureNetworkClockAvailabilityService::GetResult, "GetResult"},
+            {3, &IEnsureNetworkClockAvailabilityService::Cancel, "Cancel"},
+            {4, &IEnsureNetworkClockAvailabilityService::IsProcessing, "IsProcessing"},
+            {5, &IEnsureNetworkClockAvailabilityService::GetServerTime, "GetServerTime"},
+        };
+        RegisterHandlers(functions);
+        auto& kernel = Core::System::GetInstance().Kernel();
+        finished_event =
+            Kernel::Event::Create(kernel, Kernel::ResetType::OneShot,
+                                  "IEnsureNetworkClockAvailabilityService:FinishEvent");
+    }
+private:
+    Kernel::SharedPtr<Kernel::Event> finished_event;
+    void StartTask(Kernel::HLERequestContext& ctx) {
+        // No need to connect to the internet, just finish the task straight away.
+        finished_event->Signal();
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_DEBUG(Service_NIM, "called");
+    }
+    void GetFinishNotificationEvent(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushCopyObjects(finished_event);
+        LOG_DEBUG(Service_NIM, "called");
+    }
+    void GetResult(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_DEBUG(Service_NIM, "called");
+    }
+    void Cancel(Kernel::HLERequestContext& ctx) {
+        finished_event->Clear();
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_DEBUG(Service_NIM, "called");
+    }
+    void IsProcessing(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u32>(0); // We instantly process the request
+        LOG_DEBUG(Service_NIM, "called");
+    }
+    void GetServerTime(Kernel::HLERequestContext& ctx) {
+        const s64 server_time{std::chrono::duration_cast<std::chrono::seconds>(
+                                  std::chrono::system_clock::now().time_since_epoch())
+                                  .count()};
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<s64>(server_time);
+        LOG_DEBUG(Service_NIM, "called");
+    }
+};
 class NTC final : public ServiceFramework<NTC> {
 public:
    explicit NTC() : ServiceFramework{"ntc"} {
        // clang-format off
        static const FunctionInfo functions[] = {
-            {0, nullptr, "OpenEnsureNetworkClockAvailabilityService"},
+            {0, &NTC::OpenEnsureNetworkClockAvailabilityService, "OpenEnsureNetworkClockAvailabilityService"},
-            {100, nullptr, "SuspendAutonomicTimeCorrection"},
+            {100, &NTC::SuspendAutonomicTimeCorrection, "SuspendAutonomicTimeCorrection"},
-            {101, nullptr, "ResumeAutonomicTimeCorrection"},
+            {101, &NTC::ResumeAutonomicTimeCorrection, "ResumeAutonomicTimeCorrection"},
        };
        // clang-format on
        RegisterHandlers(functions);
    }
+private:
+    void OpenEnsureNetworkClockAvailabilityService(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IEnsureNetworkClockAvailabilityService>();
+        LOG_DEBUG(Service_NIM, "called");
+    }
+    // TODO(ogniK): Do we need these?
+    void SuspendAutonomicTimeCorrection(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_WARNING(Service_NIM, "(STUBBED) called");
+    }
+    void ResumeAutonomicTimeCorrection(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+        LOG_WARNING(Service_NIM, "(STUBBED) called");
+    }
 };
 void InstallInterfaces(SM::ServiceManager& sm) {
diff --git a/src/core/hle/service/ns/pl_u.cpp b/src/core/hle/service/ns/pl_u.cpp
index 1069d103f..4b2f758a8 100644
--- a/src/core/hle/service/ns/pl_u.cpp
+++ b/src/core/hle/service/ns/pl_u.cpp
@@ -317,9 +317,9 @@ void PL_U::GetSharedMemoryAddressOffset(Kernel::HLERequestContext& ctx) {
 void PL_U::GetSharedMemoryNativeHandle(Kernel::HLERequestContext& ctx) {
    // Map backing memory for the font data
-    Core::CurrentProcess()->vm_manager.MapMemoryBlock(SHARED_FONT_MEM_VADDR, impl->shared_font, 0,
+    Core::CurrentProcess()->VMManager().MapMemoryBlock(SHARED_FONT_MEM_VADDR, impl->shared_font, 0,
-                                                      SHARED_FONT_MEM_SIZE,
+                                                       SHARED_FONT_MEM_SIZE,
-                                                      Kernel::MemoryState::Shared);
+                                                       Kernel::MemoryState::Shared);
    // Create shared font memory object
    auto& kernel = Core::System::GetInstance().Kernel();
diff --git a/src/core/hle/service/sm/controller.cpp b/src/core/hle/service/sm/controller.cpp
index cdf328a26..98f6e4111 100644
--- a/src/core/hle/service/sm/controller.cpp
+++ b/src/core/hle/service/sm/controller.cpp
@@ -2,8 +2,11 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+#include "common/assert.h"
 #include "common/logging/log.h"
 #include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/client_session.h"
+#include "core/hle/kernel/server_session.h"
 #include "core/hle/kernel/session.h"
 #include "core/hle/service/sm/controller.h"
diff --git a/src/core/hle/service/sm/sm.cpp b/src/core/hle/service/sm/sm.cpp
index 096f0fd52..464e79d01 100644
--- a/src/core/hle/service/sm/sm.cpp
+++ b/src/core/hle/service/sm/sm.cpp
@@ -108,7 +108,7 @@ void SM::GetService(Kernel::HLERequestContext& ctx) {
    auto client_port = service_manager->GetServicePort(name);
    if (client_port.Failed()) {
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(client_port.Code());
        LOG_ERROR(Service_SM, "called service={} -> error 0x{:08X}", name, client_port.Code().raw);
        if (name.length() == 0)
@@ -121,8 +121,7 @@ void SM::GetService(Kernel::HLERequestContext& ctx) {
    ASSERT(session.Succeeded());
    if (session.Succeeded()) {
        LOG_DEBUG(Service_SM, "called service={} -> session={}", name, (*session)->GetObjectId());
-        IPC::ResponseBuilder rb =
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1, IPC::ResponseBuilder::Flags::AlwaysMoveHandles};
-            rp.MakeBuilder(2, 0, 1, IPC::ResponseBuilder::Flags::AlwaysMoveHandles);
        rb.Push(session.Code());
        rb.PushMoveObjects(std::move(session).Unwrap());
    }
diff --git a/src/core/hle/service/ssl/ssl.cpp b/src/core/hle/service/ssl/ssl.cpp
index 63b86e099..bc4f7a437 100644
--- a/src/core/hle/service/ssl/ssl.cpp
+++ b/src/core/hle/service/ssl/ssl.cpp
@@ -71,7 +71,7 @@ private:
        LOG_WARNING(Service_SSL, "(STUBBED) called");
        IPC::RequestParser rp{ctx};
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
    }
@@ -103,6 +103,7 @@ public:
    }
 private:
+    u32 ssl_version{};
    void CreateContext(Kernel::HLERequestContext& ctx) {
        LOG_WARNING(Service_SSL, "(STUBBED) called");
@@ -112,10 +113,9 @@ private:
    }
    void SetInterfaceVersion(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_SSL, "(STUBBED) called");
+        LOG_DEBUG(Service_SSL, "called");
        IPC::RequestParser rp{ctx};
-        u32 unk1 = rp.Pop<u32>(); // Probably minor/major?
+        ssl_version = rp.Pop<u32>();
-        u32 unk2 = rp.Pop<u32>(); // TODO(ogniK): Figure out what this does
        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp
index d0cde5ede..bbc02abcc 100644
--- a/src/core/hle/service/vi/vi.cpp
+++ b/src/core/hle/service/vi/vi.cpp
@@ -612,7 +612,7 @@ public:
            {3000, nullptr, "ListDisplayModes"},
            {3001, nullptr, "ListDisplayRgbRanges"},
            {3002, nullptr, "ListDisplayContentTypes"},
-            {3200, nullptr, "GetDisplayMode"},
+            {3200, &ISystemDisplayService::GetDisplayMode, "GetDisplayMode"},
            {3201, nullptr, "SetDisplayMode"},
            {3202, nullptr, "GetDisplayUnderscan"},
            {3203, nullptr, "SetDisplayUnderscan"},
@@ -650,7 +650,7 @@ private:
        u64 layer_id = rp.Pop<u64>();
        u64 z_value = rp.Pop<u64>();
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
    }
@@ -658,11 +658,29 @@ private:
        IPC::RequestParser rp{ctx};
        u64 layer_id = rp.Pop<u64>();
        bool visibility = rp.Pop<bool>();
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
        LOG_WARNING(Service_VI, "(STUBBED) called, layer_id=0x{:08X}, visibility={}", layer_id,
                    visibility);
    }
+    void GetDisplayMode(Kernel::HLERequestContext& ctx) {
+        IPC::ResponseBuilder rb{ctx, 6};
+        rb.Push(RESULT_SUCCESS);
+        if (Settings::values.use_docked_mode) {
+            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedWidth));
+            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedHeight));
+        } else {
+            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedWidth));
+            rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedHeight));
+        }
+        rb.PushRaw<float>(60.0f);
+        rb.Push<u32>(0);
+        LOG_DEBUG(Service_VI, "called");
+    }
 };
 class IManagerDisplayService final : public ServiceFramework<IManagerDisplayService> {
@@ -747,7 +765,7 @@ private:
        IPC::RequestParser rp{ctx};
        u64 display = rp.Pop<u64>();
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
    }
@@ -761,7 +779,7 @@ private:
        u64 layer_id = nv_flinger->CreateLayer(display);
-        IPC::ResponseBuilder rb = rp.MakeBuilder(4, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 4};
        rb.Push(RESULT_SUCCESS);
        rb.Push(layer_id);
    }
@@ -772,7 +790,7 @@ private:
        u32 stack = rp.Pop<u32>();
        u64 layer_id = rp.Pop<u64>();
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
    }
@@ -780,7 +798,7 @@ private:
        IPC::RequestParser rp{ctx};
        u64 layer_id = rp.Pop<u64>();
        bool visibility = rp.Pop<bool>();
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
        LOG_WARNING(Service_VI, "(STUBBED) called, layer_id=0x{:X}, visibility={}", layer_id,
                    visibility);
@@ -837,7 +855,7 @@ private:
        ASSERT_MSG(name == "Default", "Non-default displays aren't supported yet");
-        IPC::ResponseBuilder rb = rp.MakeBuilder(4, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 4};
        rb.Push(RESULT_SUCCESS);
        rb.Push<u64>(nv_flinger->OpenDisplay(name));
    }
@@ -847,7 +865,7 @@ private:
        IPC::RequestParser rp{ctx};
        u64 display_id = rp.Pop<u64>();
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
    }
@@ -856,7 +874,7 @@ private:
        IPC::RequestParser rp{ctx};
        u64 display_id = rp.Pop<u64>();
-        IPC::ResponseBuilder rb = rp.MakeBuilder(6, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 6};
        rb.Push(RESULT_SUCCESS);
        if (Settings::values.use_docked_mode) {
@@ -874,7 +892,7 @@ private:
        u32 scaling_mode = rp.Pop<u32>();
        u64 unknown = rp.Pop<u64>();
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
    }
@@ -882,7 +900,7 @@ private:
        IPC::RequestParser rp{ctx};
        DisplayInfo display_info;
        ctx.WriteBuffer(&display_info, sizeof(DisplayInfo));
-        IPC::ResponseBuilder rb = rp.MakeBuilder(4, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 4};
        rb.Push(RESULT_SUCCESS);
        rb.Push<u64>(1);
        LOG_WARNING(Service_VI, "(STUBBED) called");
@@ -903,7 +921,7 @@ private:
        u32 buffer_queue_id = nv_flinger->GetBufferQueueId(display_id, layer_id);
        NativeWindow native_window{buffer_queue_id};
-        IPC::ResponseBuilder rb = rp.MakeBuilder(4, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 4};
        rb.Push(RESULT_SUCCESS);
        rb.Push<u64>(ctx.WriteBuffer(native_window.Serialize()));
    }
@@ -922,7 +940,7 @@ private:
        u32 buffer_queue_id = nv_flinger->GetBufferQueueId(display_id, layer_id);
        NativeWindow native_window{buffer_queue_id};
-        IPC::ResponseBuilder rb = rp.MakeBuilder(6, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 6};
        rb.Push(RESULT_SUCCESS);
        rb.Push(layer_id);
        rb.Push<u64>(ctx.WriteBuffer(native_window.Serialize()));
@@ -934,7 +952,7 @@ private:
        IPC::RequestParser rp{ctx};
        u64 layer_id = rp.Pop<u64>();
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 0, 0);
+        IPC::ResponseBuilder rb{ctx, 2};
        rb.Push(RESULT_SUCCESS);
    }
@@ -945,7 +963,7 @@ private:
        auto vsync_event = nv_flinger->GetVsyncEvent(display_id);
-        IPC::ResponseBuilder rb = rp.MakeBuilder(2, 1, 0);
+        IPC::ResponseBuilder rb{ctx, 2, 1};
        rb.Push(RESULT_SUCCESS);
        rb.PushCopyObjects(vsync_event);
    }
diff --git a/src/core/loader/deconstructed_rom_directory.cpp b/src/core/loader/deconstructed_rom_directory.cpp
index 2b8f78136..c1824b9c3 100644
--- a/src/core/loader/deconstructed_rom_directory.cpp
+++ b/src/core/loader/deconstructed_rom_directory.cpp
@@ -14,11 +14,9 @@
 #include "core/gdbstub/gdbstub.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
-#include "core/hle/kernel/resource_limit.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/loader/deconstructed_rom_directory.h"
 #include "core/loader/nso.h"
-#include "core/memory.h"
 namespace Loader {
@@ -88,8 +86,7 @@ FileType AppLoader_DeconstructedRomDirectory::IdentifyType(const FileSys::Virtua
    return FileType::Error;
 }
-ResultStatus AppLoader_DeconstructedRomDirectory::Load(
+ResultStatus AppLoader_DeconstructedRomDirectory::Load(Kernel::Process& process) {
-    Kernel::SharedPtr<Kernel::Process>& process) {
    if (is_loaded) {
        return ResultStatus::ErrorAlreadyLoaded;
    }
@@ -127,12 +124,16 @@ ResultStatus AppLoader_DeconstructedRomDirectory::Load(
    metadata.Print();
    const FileSys::ProgramAddressSpaceType arch_bits{metadata.GetAddressSpaceType()};
-    if (arch_bits == FileSys::ProgramAddressSpaceType::Is32Bit) {
+    if (arch_bits == FileSys::ProgramAddressSpaceType::Is32Bit ||
+        arch_bits == FileSys::ProgramAddressSpaceType::Is32BitNoMap) {
        return ResultStatus::Error32BitISA;
    }
+    process.LoadFromMetadata(metadata);
    // Load NSO modules
-    VAddr next_load_addr{Memory::PROCESS_IMAGE_VADDR};
+    const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress();
+    VAddr next_load_addr = base_address;
    for (const auto& module : {"rtld", "main", "subsdk0", "subsdk1", "subsdk2", "subsdk3",
                               "subsdk4", "subsdk5", "subsdk6", "subsdk7", "sdk"}) {
        const FileSys::VirtualFile module_file = dir->GetFile(module);
@@ -145,13 +146,7 @@ ResultStatus AppLoader_DeconstructedRomDirectory::Load(
        }
    }
-    auto& kernel = Core::System::GetInstance().Kernel();
+    process.Run(base_address, metadata.GetMainThreadPriority(), metadata.GetMainThreadStackSize());
-    process->program_id = metadata.GetTitleID();
-    process->svc_access_mask.set();
-    process->resource_limit =
-        kernel.ResourceLimitForCategory(Kernel::ResourceLimitCategory::APPLICATION);
-    process->Run(Memory::PROCESS_IMAGE_VADDR, metadata.GetMainThreadPriority(),
-                 metadata.GetMainThreadStackSize());
    // Find the RomFS by searching for a ".romfs" file in this directory
    const auto& files = dir->GetFiles();
diff --git a/src/core/loader/deconstructed_rom_directory.h b/src/core/loader/deconstructed_rom_directory.h
index 8a0dc1b1e..d109ed2b5 100644
--- a/src/core/loader/deconstructed_rom_directory.h
+++ b/src/core/loader/deconstructed_rom_directory.h
@@ -7,7 +7,6 @@
 #include <string>
 #include "common/common_types.h"
 #include "core/file_sys/program_metadata.h"
-#include "core/hle/kernel/object.h"
 #include "core/loader/loader.h"
 namespace Loader {
@@ -38,7 +37,7 @@ public:
        return IdentifyType(file);
    }
-    ResultStatus Load(Kernel::SharedPtr<Kernel::Process>& process) override;
+    ResultStatus Load(Kernel::Process& process) override;
    ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override;
    ResultStatus ReadIcon(std::vector<u8>& buffer) override;
diff --git a/src/core/loader/elf.cpp b/src/core/loader/elf.cpp
index 0e2af20b4..e67b49fc9 100644
--- a/src/core/loader/elf.cpp
+++ b/src/core/loader/elf.cpp
@@ -12,7 +12,7 @@
 #include "core/core.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
-#include "core/hle/kernel/resource_limit.h"
+#include "core/hle/kernel/vm_manager.h"
 #include "core/loader/elf.h"
 #include "core/memory.h"
@@ -189,7 +189,7 @@ private:
    u32* sectionAddrs;
    bool relocate;
-    u32 entryPoint;
+    VAddr entryPoint;
 public:
    explicit ElfReader(void* ptr);
@@ -205,13 +205,13 @@ public:
    ElfMachine GetMachine() const {
        return (ElfMachine)(header->e_machine);
    }
-    u32 GetEntryPoint() const {
+    VAddr GetEntryPoint() const {
        return entryPoint;
    }
    u32 GetFlags() const {
        return (u32)(header->e_flags);
    }
-    SharedPtr<CodeSet> LoadInto(u32 vaddr);
+    SharedPtr<CodeSet> LoadInto(VAddr vaddr);
    int GetNumSegments() const {
        return (int)(header->e_phnum);
@@ -274,7 +274,7 @@ const char* ElfReader::GetSectionName(int section) const {
    return nullptr;
 }
-SharedPtr<CodeSet> ElfReader::LoadInto(u32 vaddr) {
+SharedPtr<CodeSet> ElfReader::LoadInto(VAddr vaddr) {
    LOG_DEBUG(Loader, "String section: {}", header->e_shstrndx);
    // Should we relocate?
@@ -289,11 +289,11 @@ SharedPtr<CodeSet> ElfReader::LoadInto(u32 vaddr) {
    LOG_DEBUG(Loader, "{} segments:", header->e_phnum);
    // First pass : Get the bits into RAM
-    u32 base_addr = relocate ? vaddr : 0;
+    const VAddr base_addr = relocate ? vaddr : 0;
-    u32 total_image_size = 0;
+    u64 total_image_size = 0;
    for (unsigned int i = 0; i < header->e_phnum; ++i) {
-        Elf32_Phdr* p = &segments[i];
+        const Elf32_Phdr* p = &segments[i];
        if (p->p_type == PT_LOAD) {
            total_image_size += (p->p_memsz + 0xFFF) & ~0xFFF;
        }
@@ -306,7 +306,7 @@ SharedPtr<CodeSet> ElfReader::LoadInto(u32 vaddr) {
    SharedPtr<CodeSet> codeset = CodeSet::Create(kernel, "");
    for (unsigned int i = 0; i < header->e_phnum; ++i) {
-        Elf32_Phdr* p = &segments[i];
+        const Elf32_Phdr* p = &segments[i];
        LOG_DEBUG(Loader, "Type: {} Vaddr: {:08X} Filesz: {:08X} Memsz: {:08X} ", p->p_type,
                  p->p_vaddr, p->p_filesz, p->p_memsz);
@@ -333,8 +333,8 @@ SharedPtr<CodeSet> ElfReader::LoadInto(u32 vaddr) {
                continue;
            }
-            u32 segment_addr = base_addr + p->p_vaddr;
+            const VAddr segment_addr = base_addr + p->p_vaddr;
-            u32 aligned_size = (p->p_memsz + 0xFFF) & ~0xFFF;
+            const u32 aligned_size = (p->p_memsz + 0xFFF) & ~0xFFF;
            codeset_segment->offset = current_image_position;
            codeset_segment->addr = segment_addr;
@@ -387,7 +387,7 @@ FileType AppLoader_ELF::IdentifyType(const FileSys::VirtualFile& file) {
    return FileType::Error;
 }
-ResultStatus AppLoader_ELF::Load(Kernel::SharedPtr<Kernel::Process>& process) {
+ResultStatus AppLoader_ELF::Load(Kernel::Process& process) {
    if (is_loaded)
        return ResultStatus::ErrorAlreadyLoaded;
@@ -395,19 +395,13 @@ ResultStatus AppLoader_ELF::Load(Kernel::SharedPtr<Kernel::Process>& process) {
    if (buffer.size() != file->GetSize())
        return ResultStatus::ErrorIncorrectELFFileSize;
+    const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress();
    ElfReader elf_reader(&buffer[0]);
-    SharedPtr<CodeSet> codeset = elf_reader.LoadInto(Memory::PROCESS_IMAGE_VADDR);
+    SharedPtr<CodeSet> codeset = elf_reader.LoadInto(base_address);
    codeset->name = file->GetName();
-    process->LoadModule(codeset, codeset->entrypoint);
+    process.LoadModule(codeset, codeset->entrypoint);
-    process->svc_access_mask.set();
+    process.Run(codeset->entrypoint, 48, Memory::DEFAULT_STACK_SIZE);
-    // Attach the default resource limit (APPLICATION) to the process
-    auto& kernel = Core::System::GetInstance().Kernel();
-    process->resource_limit =
-        kernel.ResourceLimitForCategory(Kernel::ResourceLimitCategory::APPLICATION);
-    process->Run(codeset->entrypoint, 48, Memory::DEFAULT_STACK_SIZE);
    is_loaded = true;
    return ResultStatus::Success;
diff --git a/src/core/loader/elf.h b/src/core/loader/elf.h
index b8fb982d0..6af76441c 100644
--- a/src/core/loader/elf.h
+++ b/src/core/loader/elf.h
@@ -8,9 +8,6 @@
 #include "common/common_types.h"
 #include "core/loader/loader.h"
-////////////////////////////////////////////////////////////////////////////////////////////////////
-// Loader namespace
 namespace Loader {
 /// Loads an ELF/AXF file
@@ -29,7 +26,7 @@ public:
        return IdentifyType(file);
    }
-    ResultStatus Load(Kernel::SharedPtr<Kernel::Process>& process) override;
+    ResultStatus Load(Kernel::Process& process) override;
 };
 } // namespace Loader
diff --git a/src/core/loader/loader.h b/src/core/loader/loader.h
index 843c4bb91..20e66109b 100644
--- a/src/core/loader/loader.h
+++ b/src/core/loader/loader.h
@@ -12,7 +12,6 @@
 #include <boost/optional.hpp>
 #include "common/common_types.h"
 #include "core/file_sys/vfs.h"
-#include "core/hle/kernel/object.h"
 namespace Kernel {
 struct AddressMapping;
@@ -136,7 +135,7 @@ public:
     * @param process The newly created process.
     * @return The status result of the operation.
     */
-    virtual ResultStatus Load(Kernel::SharedPtr<Kernel::Process>& process) = 0;
+    virtual ResultStatus Load(Kernel::Process& process) = 0;
    /**
     * Loads the system mode that this application needs.
diff --git a/src/core/loader/nax.cpp b/src/core/loader/nax.cpp
index b46d81c02..073fb9d2f 100644
--- a/src/core/loader/nax.cpp
+++ b/src/core/loader/nax.cpp
@@ -11,6 +11,20 @@
 #include "core/loader/nca.h"
 namespace Loader {
+namespace {
+FileType IdentifyTypeImpl(const FileSys::NAX& nax) {
+    if (nax.GetStatus() != ResultStatus::Success) {
+        return FileType::Error;
+    }
+    const auto nca = nax.AsNCA();
+    if (nca == nullptr || nca->GetStatus() != ResultStatus::Success) {
+        return FileType::Error;
+    }
+    return FileType::NAX;
+}
+} // Anonymous namespace
 AppLoader_NAX::AppLoader_NAX(FileSys::VirtualFile file)
    : AppLoader(file), nax(std::make_unique<FileSys::NAX>(file)),
@@ -19,17 +33,15 @@ AppLoader_NAX::AppLoader_NAX(FileSys::VirtualFile file)
 AppLoader_NAX::~AppLoader_NAX() = default;
 FileType AppLoader_NAX::IdentifyType(const FileSys::VirtualFile& file) {
-    FileSys::NAX nax(file);
+    const FileSys::NAX nax(file);
+    return IdentifyTypeImpl(nax);
-    if (nax.GetStatus() == ResultStatus::Success && nax.AsNCA() != nullptr &&
+}
-        nax.AsNCA()->GetStatus() == ResultStatus::Success) {
-        return FileType::NAX;
-    }
-    return FileType::Error;
+FileType AppLoader_NAX::GetFileType() {
+    return IdentifyTypeImpl(*nax);
 }
-ResultStatus AppLoader_NAX::Load(Kernel::SharedPtr<Kernel::Process>& process) {
+ResultStatus AppLoader_NAX::Load(Kernel::Process& process) {
    if (is_loaded) {
        return ResultStatus::ErrorAlreadyLoaded;
    }
diff --git a/src/core/loader/nax.h b/src/core/loader/nax.h
index 4dbae2918..fc3c01876 100644
--- a/src/core/loader/nax.h
+++ b/src/core/loader/nax.h
@@ -31,11 +31,9 @@ public:
     */
    static FileType IdentifyType(const FileSys::VirtualFile& file);
-    FileType GetFileType() override {
+    FileType GetFileType() override;
-        return IdentifyType(file);
-    }
-    ResultStatus Load(Kernel::SharedPtr<Kernel::Process>& process) override;
+    ResultStatus Load(Kernel::Process& process) override;
    ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override;
    ResultStatus ReadProgramId(u64& out_program_id) override;
diff --git a/src/core/loader/nca.cpp b/src/core/loader/nca.cpp
index 6aaffae59..7e1b0d84f 100644
--- a/src/core/loader/nca.cpp
+++ b/src/core/loader/nca.cpp
@@ -30,7 +30,7 @@ FileType AppLoader_NCA::IdentifyType(const FileSys::VirtualFile& file) {
    return FileType::Error;
 }
-ResultStatus AppLoader_NCA::Load(Kernel::SharedPtr<Kernel::Process>& process) {
+ResultStatus AppLoader_NCA::Load(Kernel::Process& process) {
    if (is_loaded) {
        return ResultStatus::ErrorAlreadyLoaded;
    }
diff --git a/src/core/loader/nca.h b/src/core/loader/nca.h
index 10be197c4..95d9b73a1 100644
--- a/src/core/loader/nca.h
+++ b/src/core/loader/nca.h
@@ -6,7 +6,6 @@
 #include "common/common_types.h"
 #include "core/file_sys/vfs.h"
-#include "core/hle/kernel/object.h"
 #include "core/loader/loader.h"
 namespace FileSys {
@@ -34,7 +33,7 @@ public:
        return IdentifyType(file);
    }
-    ResultStatus Load(Kernel::SharedPtr<Kernel::Process>& process) override;
+    ResultStatus Load(Kernel::Process& process) override;
    ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override;
    u64 ReadRomFSIVFCOffset() const override;
diff --git a/src/core/loader/nro.cpp b/src/core/loader/nro.cpp
index c49ec34ab..c10f826a4 100644
--- a/src/core/loader/nro.cpp
+++ b/src/core/loader/nro.cpp
@@ -16,7 +16,7 @@
 #include "core/gdbstub/gdbstub.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
-#include "core/hle/kernel/resource_limit.h"
+#include "core/hle/kernel/vm_manager.h"
 #include "core/loader/nro.h"
 #include "core/memory.h"
@@ -175,23 +175,19 @@ bool AppLoader_NRO::LoadNro(FileSys::VirtualFile file, VAddr load_base) {
    return true;
 }
-ResultStatus AppLoader_NRO::Load(Kernel::SharedPtr<Kernel::Process>& process) {
+ResultStatus AppLoader_NRO::Load(Kernel::Process& process) {
    if (is_loaded) {
        return ResultStatus::ErrorAlreadyLoaded;
    }
    // Load NRO
-    static constexpr VAddr base_addr{Memory::PROCESS_IMAGE_VADDR};
+    const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress();
-    if (!LoadNro(file, base_addr)) {
+    if (!LoadNro(file, base_address)) {
        return ResultStatus::ErrorLoadingNRO;
    }
-    auto& kernel = Core::System::GetInstance().Kernel();
+    process.Run(base_address, Kernel::THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE);
-    process->svc_access_mask.set();
-    process->resource_limit =
-        kernel.ResourceLimitForCategory(Kernel::ResourceLimitCategory::APPLICATION);
-    process->Run(base_addr, Kernel::THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE);
    is_loaded = true;
    return ResultStatus::Success;
diff --git a/src/core/loader/nro.h b/src/core/loader/nro.h
index 96d2de305..04b46119a 100644
--- a/src/core/loader/nro.h
+++ b/src/core/loader/nro.h
@@ -6,7 +6,6 @@
 #include <string>
 #include "common/common_types.h"
-#include "core/hle/kernel/object.h"
 #include "core/loader/linker.h"
 #include "core/loader/loader.h"
@@ -33,7 +32,7 @@ public:
        return IdentifyType(file);
    }
-    ResultStatus Load(Kernel::SharedPtr<Kernel::Process>& process) override;
+    ResultStatus Load(Kernel::Process& process) override;
    ResultStatus ReadIcon(std::vector<u8>& buffer) override;
    ResultStatus ReadProgramId(u64& out_program_id) override;
diff --git a/src/core/loader/nso.cpp b/src/core/loader/nso.cpp
index 3c6306818..cbe2a3e53 100644
--- a/src/core/loader/nso.cpp
+++ b/src/core/loader/nso.cpp
@@ -13,7 +13,7 @@
 #include "core/gdbstub/gdbstub.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
-#include "core/hle/kernel/resource_limit.h"
+#include "core/hle/kernel/vm_manager.h"
 #include "core/loader/nso.h"
 #include "core/memory.h"
@@ -32,11 +32,18 @@ static_assert(sizeof(NsoSegmentHeader) == 0x10, "NsoSegmentHeader has incorrect
 struct NsoHeader {
    u32_le magic;
-    INSERT_PADDING_BYTES(0xc);
+    u32_le version;
+    INSERT_PADDING_WORDS(1);
+    u8 flags;
    std::array<NsoSegmentHeader, 3> segments; // Text, RoData, Data (in that order)
    u32_le bss_size;
    INSERT_PADDING_BYTES(0x1c);
    std::array<u32_le, 3> segments_compressed_size;
+    bool IsSegmentCompressed(size_t segment_num) const {
+        ASSERT_MSG(segment_num < 3, "Invalid segment {}", segment_num);
+        return ((flags >> segment_num) & 1);
+    }
 };
 static_assert(sizeof(NsoHeader) == 0x6c, "NsoHeader has incorrect size.");
 static_assert(std::is_trivially_copyable_v<NsoHeader>, "NsoHeader isn't trivially copyable.");
@@ -105,9 +112,11 @@ VAddr AppLoader_NSO::LoadModule(FileSys::VirtualFile file, VAddr load_base) {
    Kernel::SharedPtr<Kernel::CodeSet> codeset = Kernel::CodeSet::Create(kernel, "");
    std::vector<u8> program_image;
    for (std::size_t i = 0; i < nso_header.segments.size(); ++i) {
-        const std::vector<u8> compressed_data =
+        std::vector<u8> data =
            file->ReadBytes(nso_header.segments_compressed_size[i], nso_header.segments[i].offset);
-        std::vector<u8> data = DecompressSegment(compressed_data, nso_header.segments[i]);
+        if (nso_header.IsSegmentCompressed(i)) {
+            data = DecompressSegment(data, nso_header.segments[i]);
+        }
        program_image.resize(nso_header.segments[i].location);
        program_image.insert(program_image.end(), data.begin(), data.end());
        codeset->segments[i].addr = nso_header.segments[i].location;
@@ -144,21 +153,17 @@ VAddr AppLoader_NSO::LoadModule(FileSys::VirtualFile file, VAddr load_base) {
    return load_base + image_size;
 }
-ResultStatus AppLoader_NSO::Load(Kernel::SharedPtr<Kernel::Process>& process) {
+ResultStatus AppLoader_NSO::Load(Kernel::Process& process) {
    if (is_loaded) {
        return ResultStatus::ErrorAlreadyLoaded;
    }
    // Load module
-    LoadModule(file, Memory::PROCESS_IMAGE_VADDR);
+    const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress();
-    LOG_DEBUG(Loader, "loaded module {} @ 0x{:X}", file->GetName(), Memory::PROCESS_IMAGE_VADDR);
+    LoadModule(file, base_address);
+    LOG_DEBUG(Loader, "loaded module {} @ 0x{:X}", file->GetName(), base_address);
-    auto& kernel = Core::System::GetInstance().Kernel();
+    process.Run(base_address, Kernel::THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE);
-    process->svc_access_mask.set();
-    process->resource_limit =
-        kernel.ResourceLimitForCategory(Kernel::ResourceLimitCategory::APPLICATION);
-    process->Run(Memory::PROCESS_IMAGE_VADDR, Kernel::THREADPRIO_DEFAULT,
-                 Memory::DEFAULT_STACK_SIZE);
    is_loaded = true;
    return ResultStatus::Success;
diff --git a/src/core/loader/nso.h b/src/core/loader/nso.h
index aaeb1f2a9..7f142405b 100644
--- a/src/core/loader/nso.h
+++ b/src/core/loader/nso.h
@@ -4,9 +4,7 @@
 #pragma once
-#include <string>
 #include "common/common_types.h"
-#include "core/hle/kernel/object.h"
 #include "core/loader/linker.h"
 #include "core/loader/loader.h"
@@ -30,7 +28,7 @@ public:
    static VAddr LoadModule(FileSys::VirtualFile file, VAddr load_base);
-    ResultStatus Load(Kernel::SharedPtr<Kernel::Process>& process) override;
+    ResultStatus Load(Kernel::Process& process) override;
 };
 } // namespace Loader
diff --git a/src/core/loader/nsp.cpp b/src/core/loader/nsp.cpp
index 291a9876d..b7ba77ef4 100644
--- a/src/core/loader/nsp.cpp
+++ b/src/core/loader/nsp.cpp
@@ -10,8 +10,6 @@
 #include "core/file_sys/control_metadata.h"
 #include "core/file_sys/nca_metadata.h"
 #include "core/file_sys/patch_manager.h"
-#include "core/file_sys/registered_cache.h"
-#include "core/file_sys/romfs.h"
 #include "core/file_sys/submission_package.h"
 #include "core/hle/kernel/process.h"
 #include "core/loader/deconstructed_rom_directory.h"
@@ -62,7 +60,7 @@ FileType AppLoader_NSP::IdentifyType(const FileSys::VirtualFile& file) {
    return FileType::Error;
 }
-ResultStatus AppLoader_NSP::Load(Kernel::SharedPtr<Kernel::Process>& process) {
+ResultStatus AppLoader_NSP::Load(Kernel::Process& process) {
    if (is_loaded) {
        return ResultStatus::ErrorAlreadyLoaded;
    }
diff --git a/src/core/loader/nsp.h b/src/core/loader/nsp.h
index 7ef810499..eac9b819a 100644
--- a/src/core/loader/nsp.h
+++ b/src/core/loader/nsp.h
@@ -35,7 +35,7 @@ public:
        return IdentifyType(file);
    }
-    ResultStatus Load(Kernel::SharedPtr<Kernel::Process>& process) override;
+    ResultStatus Load(Kernel::Process& process) override;
    ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override;
    ResultStatus ReadProgramId(u64& out_program_id) override;
diff --git a/src/core/loader/xci.cpp b/src/core/loader/xci.cpp
index 16509229f..eda67a8c8 100644
--- a/src/core/loader/xci.cpp
+++ b/src/core/loader/xci.cpp
@@ -9,8 +9,6 @@
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/control_metadata.h"
 #include "core/file_sys/patch_manager.h"
-#include "core/file_sys/romfs.h"
-#include "core/file_sys/submission_package.h"
 #include "core/hle/kernel/process.h"
 #include "core/loader/nca.h"
 #include "core/loader/xci.h"
@@ -46,7 +44,7 @@ FileType AppLoader_XCI::IdentifyType(const FileSys::VirtualFile& file) {
    return FileType::Error;
 }
-ResultStatus AppLoader_XCI::Load(Kernel::SharedPtr<Kernel::Process>& process) {
+ResultStatus AppLoader_XCI::Load(Kernel::Process& process) {
    if (is_loaded) {
        return ResultStatus::ErrorAlreadyLoaded;
    }
diff --git a/src/core/loader/xci.h b/src/core/loader/xci.h
index cc4287e17..17e47b658 100644
--- a/src/core/loader/xci.h
+++ b/src/core/loader/xci.h
@@ -35,7 +35,7 @@ public:
        return IdentifyType(file);
    }
-    ResultStatus Load(Kernel::SharedPtr<Kernel::Process>& process) override;
+    ResultStatus Load(Kernel::Process& process) override;
    ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override;
    ResultStatus ReadProgramId(u64& out_program_id) override;
diff --git a/src/core/memory.cpp b/src/core/memory.cpp
index 316b46820..014298ed6 100644
--- a/src/core/memory.cpp
+++ b/src/core/memory.cpp
@@ -3,7 +3,6 @@
 // Refer to the license.txt file included.
 #include <algorithm>
-#include <array>
 #include <cstring>
 #include <utility>
@@ -15,11 +14,11 @@
 #include "core/arm/arm_interface.h"
 #include "core/core.h"
 #include "core/hle/kernel/process.h"
+#include "core/hle/kernel/vm_manager.h"
 #include "core/hle/lock.h"
 #include "core/memory.h"
 #include "core/memory_setup.h"
 #include "video_core/renderer_base.h"
-#include "video_core/video_core.h"
 namespace Memory {
@@ -41,6 +40,21 @@ PageTable* GetCurrentPageTable() {
    return current_page_table;
 }
+PageTable::PageTable() = default;
+PageTable::PageTable(std::size_t address_space_width_in_bits) {
+    Resize(address_space_width_in_bits);
+}
+PageTable::~PageTable() = default;
+void PageTable::Resize(std::size_t address_space_width_in_bits) {
+    const std::size_t num_page_table_entries = 1ULL << (address_space_width_in_bits - PAGE_BITS);
+    pointers.resize(num_page_table_entries);
+    attributes.resize(num_page_table_entries);
+}
 static void MapPages(PageTable& page_table, VAddr base, u64 size, u8* memory, PageType type) {
    LOG_DEBUG(HW_Memory, "Mapping {} onto {:016X}-{:016X}", fmt::ptr(memory), base * PAGE_SIZE,
              (base + size) * PAGE_SIZE);
@@ -50,7 +64,7 @@ static void MapPages(PageTable& page_table, VAddr base, u64 size, u8* memory, Pa
    VAddr end = base + size;
    while (base != end) {
-        ASSERT_MSG(base < PAGE_TABLE_NUM_ENTRIES, "out of range mapping at {:016X}", base);
+        ASSERT_MSG(base < page_table.pointers.size(), "out of range mapping at {:016X}", base);
        page_table.attributes[base] = type;
        page_table.pointers[base] = memory;
@@ -105,7 +119,7 @@ void RemoveDebugHook(PageTable& page_table, VAddr base, u64 size, MemoryHookPoin
 static u8* GetPointerFromVMA(const Kernel::Process& process, VAddr vaddr) {
    u8* direct_pointer = nullptr;
-    auto& vm_manager = process.vm_manager;
+    auto& vm_manager = process.VMManager();
    auto it = vm_manager.FindVMA(vaddr);
    ASSERT(it != vm_manager.vma_map.end());
@@ -200,7 +214,7 @@ void Write(const VAddr vaddr, const T data) {
 }
 bool IsValidVirtualAddress(const Kernel::Process& process, const VAddr vaddr) {
-    auto& page_table = process.vm_manager.page_table;
+    const auto& page_table = process.VMManager().page_table;
    const u8* page_pointer = page_table.pointers[vaddr >> PAGE_BITS];
    if (page_pointer)
@@ -323,7 +337,7 @@ void RasterizerFlushVirtualRegion(VAddr start, u64 size, FlushMode mode) {
        return;
    }
-    VAddr end = start + size;
+    const VAddr end = start + size;
    const auto CheckRegion = [&](VAddr region_start, VAddr region_end) {
        if (start >= region_end || end <= region_start) {
@@ -333,7 +347,7 @@ void RasterizerFlushVirtualRegion(VAddr start, u64 size, FlushMode mode) {
        const VAddr overlap_start = std::max(start, region_start);
        const VAddr overlap_end = std::min(end, region_end);
-        const u64 overlap_size = overlap_end - overlap_start;
+        const VAddr overlap_size = overlap_end - overlap_start;
        auto& rasterizer = system_instance.Renderer().Rasterizer();
        switch (mode) {
@@ -349,8 +363,10 @@ void RasterizerFlushVirtualRegion(VAddr start, u64 size, FlushMode mode) {
        }
    };
-    CheckRegion(PROCESS_IMAGE_VADDR, PROCESS_IMAGE_VADDR_END);
+    const auto& vm_manager = Core::CurrentProcess()->VMManager();
-    CheckRegion(HEAP_VADDR, HEAP_VADDR_END);
+    CheckRegion(vm_manager.GetCodeRegionBaseAddress(), vm_manager.GetCodeRegionEndAddress());
+    CheckRegion(vm_manager.GetHeapRegionBaseAddress(), vm_manager.GetHeapRegionEndAddress());
 }
 u8 Read8(const VAddr addr) {
@@ -371,7 +387,7 @@ u64 Read64(const VAddr addr) {
 void ReadBlock(const Kernel::Process& process, const VAddr src_addr, void* dest_buffer,
               const std::size_t size) {
-    auto& page_table = process.vm_manager.page_table;
+    const auto& page_table = process.VMManager().page_table;
    std::size_t remaining_size = size;
    std::size_t page_index = src_addr >> PAGE_BITS;
@@ -436,7 +452,7 @@ void Write64(const VAddr addr, const u64 data) {
 void WriteBlock(const Kernel::Process& process, const VAddr dest_addr, const void* src_buffer,
                const std::size_t size) {
-    auto& page_table = process.vm_manager.page_table;
+    const auto& page_table = process.VMManager().page_table;
    std::size_t remaining_size = size;
    std::size_t page_index = dest_addr >> PAGE_BITS;
    std::size_t page_offset = dest_addr & PAGE_MASK;
@@ -482,7 +498,7 @@ void WriteBlock(const VAddr dest_addr, const void* src_buffer, const std::size_t
 }
 void ZeroBlock(const Kernel::Process& process, const VAddr dest_addr, const std::size_t size) {
-    auto& page_table = process.vm_manager.page_table;
+    const auto& page_table = process.VMManager().page_table;
    std::size_t remaining_size = size;
    std::size_t page_index = dest_addr >> PAGE_BITS;
    std::size_t page_offset = dest_addr & PAGE_MASK;
@@ -524,7 +540,7 @@ void ZeroBlock(const Kernel::Process& process, const VAddr dest_addr, const std:
 void CopyBlock(const Kernel::Process& process, VAddr dest_addr, VAddr src_addr,
               const std::size_t size) {
-    auto& page_table = process.vm_manager.page_table;
+    const auto& page_table = process.VMManager().page_table;
    std::size_t remaining_size = size;
    std::size_t page_index = src_addr >> PAGE_BITS;
    std::size_t page_offset = src_addr & PAGE_MASK;
diff --git a/src/core/memory.h b/src/core/memory.h
index 2a27c0251..1acf5ce8c 100644
--- a/src/core/memory.h
+++ b/src/core/memory.h
@@ -4,10 +4,10 @@
 #pragma once
-#include <array>
 #include <cstddef>
 #include <string>
 #include <tuple>
+#include <vector>
 #include <boost/icl/interval_map.hpp>
 #include "common/common_types.h"
 #include "core/memory_hook.h"
@@ -23,10 +23,8 @@ namespace Memory {
 * be mapped.
 */
 constexpr std::size_t PAGE_BITS = 12;
-constexpr u64 PAGE_SIZE = 1 << PAGE_BITS;
+constexpr u64 PAGE_SIZE = 1ULL << PAGE_BITS;
 constexpr u64 PAGE_MASK = PAGE_SIZE - 1;
-constexpr std::size_t ADDRESS_SPACE_BITS = 36;
-constexpr std::size_t PAGE_TABLE_NUM_ENTRIES = 1ULL << (ADDRESS_SPACE_BITS - PAGE_BITS);
 enum class PageType : u8 {
    /// Page is unmapped and should cause an access error.
@@ -62,32 +60,39 @@ struct SpecialRegion {
 * mimics the way a real CPU page table works.
 */
 struct PageTable {
+    explicit PageTable();
+    explicit PageTable(std::size_t address_space_width_in_bits);
+    ~PageTable();
+    /**
+     * Resizes the page table to be able to accomodate enough pages within
+     * a given address space.
+     *
+     * @param address_space_width_in_bits The address size width in bits.
+     */
+    void Resize(std::size_t address_space_width_in_bits);
    /**
-     * Array of memory pointers backing each page. An entry can only be non-null if the
+     * Vector of memory pointers backing each page. An entry can only be non-null if the
-     * corresponding entry in the `attributes` array is of type `Memory`.
+     * corresponding entry in the `attributes` vector is of type `Memory`.
     */
-    std::array<u8*, PAGE_TABLE_NUM_ENTRIES> pointers;
+    std::vector<u8*> pointers;
    /**
-     * Contains MMIO handlers that back memory regions whose entries in the `attribute` array is of
+     * Contains MMIO handlers that back memory regions whose entries in the `attribute` vector is
-     * type `Special`.
+     * of type `Special`.
     */
    boost::icl::interval_map<VAddr, std::set<SpecialRegion>> special_regions;
    /**
-     * Array of fine grained page attributes. If it is set to any value other than `Memory`, then
+     * Vector of fine grained page attributes. If it is set to any value other than `Memory`, then
     * the corresponding entry in `pointers` MUST be set to null.
     */
-    std::array<PageType, PAGE_TABLE_NUM_ENTRIES> attributes;
+    std::vector<PageType> attributes;
 };
 /// Virtual user-space memory regions
 enum : VAddr {
-    /// Where the application text, data and bss reside.
-    PROCESS_IMAGE_VADDR = 0x08000000,
-    PROCESS_IMAGE_MAX_SIZE = 0x08000000,
-    PROCESS_IMAGE_VADDR_END = PROCESS_IMAGE_VADDR + PROCESS_IMAGE_MAX_SIZE,
    /// Read-only page containing kernel and system configuration values.
    CONFIG_MEMORY_VADDR = 0x1FF80000,
    CONFIG_MEMORY_SIZE = 0x00001000,
@@ -98,36 +103,12 @@ enum : VAddr {
    SHARED_PAGE_SIZE = 0x00001000,
    SHARED_PAGE_VADDR_END = SHARED_PAGE_VADDR + SHARED_PAGE_SIZE,
-    /// Area where TLS (Thread-Local Storage) buffers are allocated.
+    /// TLS (Thread-Local Storage) related.
-    TLS_AREA_VADDR = 0x40000000,
    TLS_ENTRY_SIZE = 0x200,
-    TLS_AREA_SIZE = 0x10000000,
-    TLS_AREA_VADDR_END = TLS_AREA_VADDR + TLS_AREA_SIZE,
    /// Application stack
-    STACK_AREA_VADDR = TLS_AREA_VADDR_END,
-    STACK_AREA_SIZE = 0x10000000,
-    STACK_AREA_VADDR_END = STACK_AREA_VADDR + STACK_AREA_SIZE,
    DEFAULT_STACK_SIZE = 0x100000,
-    /// Application heap
-    /// Size is confirmed to be a static value on fw 3.0.0
-    HEAP_VADDR = 0x108000000,
-    HEAP_SIZE = 0x180000000,
-    HEAP_VADDR_END = HEAP_VADDR + HEAP_SIZE,
-    /// New map region
-    /// Size is confirmed to be a static value on fw 3.0.0
-    NEW_MAP_REGION_VADDR = HEAP_VADDR_END,
-    NEW_MAP_REGION_SIZE = 0x80000000,
-    NEW_MAP_REGION_VADDR_END = NEW_MAP_REGION_VADDR + NEW_MAP_REGION_SIZE,
-    /// Map region
-    /// Size is confirmed to be a static value on fw 3.0.0
-    MAP_REGION_VADDR = NEW_MAP_REGION_VADDR_END,
-    MAP_REGION_SIZE = 0x1000000000,
-    MAP_REGION_VADDR_END = MAP_REGION_VADDR + MAP_REGION_SIZE,
    /// Kernel Virtual Address Range
    KERNEL_REGION_VADDR = 0xFFFFFF8000000000,
    KERNEL_REGION_SIZE = 0x7FFFE00000,
diff --git a/src/tests/core/arm/arm_test_common.cpp b/src/tests/core/arm/arm_test_common.cpp
index 7c69fc26e..c0a57e71f 100644
--- a/src/tests/core/arm/arm_test_common.cpp
+++ b/src/tests/core/arm/arm_test_common.cpp
@@ -2,6 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+#include <algorithm>
 #include "core/core.h"
 #include "core/hle/kernel/process.h"
 #include "core/memory.h"
@@ -14,11 +16,12 @@ TestEnvironment::TestEnvironment(bool mutable_memory_)
    : mutable_memory(mutable_memory_), test_memory(std::make_shared<TestMemory>(this)) {
    Core::CurrentProcess() = Kernel::Process::Create(kernel, "");
-    page_table = &Core::CurrentProcess()->vm_manager.page_table;
+    page_table = &Core::CurrentProcess()->VMManager().page_table;
-    page_table->pointers.fill(nullptr);
+    std::fill(page_table->pointers.begin(), page_table->pointers.end(), nullptr);
    page_table->special_regions.clear();
-    page_table->attributes.fill(Memory::PageType::Unmapped);
+    std::fill(page_table->attributes.begin(), page_table->attributes.end(),
+              Memory::PageType::Unmapped);
    Memory::MapIoRegion(*page_table, 0x00000000, 0x80000000, test_memory);
    Memory::MapIoRegion(*page_table, 0x80000000, 0x80000000, test_memory);
diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h
index b81b0723d..9f5581045 100644
--- a/src/video_core/engines/maxwell_3d.h
+++ b/src/video_core/engines/maxwell_3d.h
@@ -41,6 +41,7 @@ public:
        static constexpr std::size_t NumCBData = 16;
        static constexpr std::size_t NumVertexArrays = 32;
        static constexpr std::size_t NumVertexAttributes = 32;
+        static constexpr std::size_t NumTextureSamplers = 32;
        static constexpr std::size_t MaxShaderProgram = 6;
        static constexpr std::size_t MaxShaderStage = 5;
        // Maximum number of const buffers per shader stage.
@@ -461,7 +462,11 @@ public:
                    u32 entry;
                } macros;
-                INSERT_PADDING_WORDS(0x1B8);
+                INSERT_PADDING_WORDS(0x189);
+                u32 tfb_enabled;
+                INSERT_PADDING_WORDS(0x2E);
                RenderTargetConfig rt[NumRenderTargets];
@@ -594,7 +599,9 @@ public:
                u32 depth_write_enabled;
-                INSERT_PADDING_WORDS(0x7);
+                u32 alpha_test_enabled;
+                INSERT_PADDING_WORDS(0x6);
                u32 d3d_cull_mode;
@@ -635,7 +642,11 @@ public:
                u32 vb_element_base;
-                INSERT_PADDING_WORDS(0x40);
+                INSERT_PADDING_WORDS(0x38);
+                float point_size;
+                INSERT_PADDING_WORDS(0x7);
                u32 zeta_enable;
@@ -977,6 +988,7 @@ private:
                  "Field " #field_name " has invalid position")
 ASSERT_REG_POSITION(macros, 0x45);
+ASSERT_REG_POSITION(tfb_enabled, 0x1D1);
 ASSERT_REG_POSITION(rt, 0x200);
 ASSERT_REG_POSITION(viewport_transform[0], 0x280);
 ASSERT_REG_POSITION(viewport, 0x300);
@@ -996,6 +1008,7 @@ ASSERT_REG_POSITION(zeta_height, 0x48b);
 ASSERT_REG_POSITION(depth_test_enable, 0x4B3);
 ASSERT_REG_POSITION(independent_blend_enable, 0x4B9);
 ASSERT_REG_POSITION(depth_write_enabled, 0x4BA);
+ASSERT_REG_POSITION(alpha_test_enabled, 0x4BB);
 ASSERT_REG_POSITION(d3d_cull_mode, 0x4C2);
 ASSERT_REG_POSITION(depth_test_func, 0x4C3);
 ASSERT_REG_POSITION(blend, 0x4CF);
@@ -1009,6 +1022,7 @@ ASSERT_REG_POSITION(stencil_front_func_mask, 0x4E6);
 ASSERT_REG_POSITION(stencil_front_mask, 0x4E7);
 ASSERT_REG_POSITION(screen_y_control, 0x4EB);
 ASSERT_REG_POSITION(vb_element_base, 0x50D);
+ASSERT_REG_POSITION(point_size, 0x546);
 ASSERT_REG_POSITION(zeta_enable, 0x54E);
 ASSERT_REG_POSITION(tsc, 0x557);
 ASSERT_REG_POSITION(tic, 0x55D);
diff --git a/src/video_core/engines/maxwell_compute.cpp b/src/video_core/engines/maxwell_compute.cpp
index e4e5f9e5e..59e28b22d 100644
--- a/src/video_core/engines/maxwell_compute.cpp
+++ b/src/video_core/engines/maxwell_compute.cpp
@@ -2,12 +2,29 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+#include "common/logging/log.h"
+#include "core/core.h"
 #include "video_core/engines/maxwell_compute.h"
 namespace Tegra {
 namespace Engines {
-void MaxwellCompute::WriteReg(u32 method, u32 value) {}
+void MaxwellCompute::WriteReg(u32 method, u32 value) {
+    ASSERT_MSG(method < Regs::NUM_REGS,
+               "Invalid MaxwellCompute register, increase the size of the Regs structure");
+    regs.reg_array[method] = value;
+    switch (method) {
+    case MAXWELL_COMPUTE_REG_INDEX(compute): {
+        LOG_CRITICAL(HW_GPU, "Compute shaders are not implemented");
+        UNREACHABLE();
+        break;
+    }
+    default:
+        break;
+    }
+}
 } // namespace Engines
 } // namespace Tegra
diff --git a/src/video_core/engines/maxwell_compute.h b/src/video_core/engines/maxwell_compute.h
index 2b3e4ced6..6ea934fb9 100644
--- a/src/video_core/engines/maxwell_compute.h
+++ b/src/video_core/engines/maxwell_compute.h
@@ -4,17 +4,53 @@
 #pragma once
+#include <array>
+#include "common/assert.h"
+#include "common/bit_field.h"
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 namespace Tegra::Engines {
+#define MAXWELL_COMPUTE_REG_INDEX(field_name)                                                      \
+    (offsetof(Tegra::Engines::MaxwellCompute::Regs, field_name) / sizeof(u32))
 class MaxwellCompute final {
 public:
    MaxwellCompute() = default;
    ~MaxwellCompute() = default;
+    struct Regs {
+        static constexpr std::size_t NUM_REGS = 0xCF8;
+        union {
+            struct {
+                INSERT_PADDING_WORDS(0x281);
+                union {
+                    u32 compute_end;
+                    BitField<0, 1, u32> unknown;
+                } compute;
+                INSERT_PADDING_WORDS(0xA76);
+            };
+            std::array<u32, NUM_REGS> reg_array;
+        };
+    } regs{};
+    static_assert(sizeof(Regs) == Regs::NUM_REGS * sizeof(u32),
+                  "MaxwellCompute Regs has wrong size");
    /// Write the value to the register identified by method.
    void WriteReg(u32 method, u32 value);
 };
+#define ASSERT_REG_POSITION(field_name, position)                                                  \
+    static_assert(offsetof(MaxwellCompute::Regs, field_name) == position * 4,                      \
+                  "Field " #field_name " has invalid position")
+ASSERT_REG_POSITION(compute, 0x281);
+#undef ASSERT_REG_POSITION
 } // namespace Tegra::Engines
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index 7e1de0fa1..b1f137b9c 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -5,9 +5,8 @@
 #pragma once
 #include <bitset>
-#include <cstring>
-#include <map>
 #include <string>
+#include <tuple>
 #include <vector>
 #include <boost/optional.hpp>
@@ -315,17 +314,29 @@ enum class TextureMiscMode : u64 {
    PTP,
 };
-enum class IpaInterpMode : u64 { Linear = 0, Perspective = 1, Flat = 2, Sc = 3 };
+enum class IpaInterpMode : u64 {
-enum class IpaSampleMode : u64 { Default = 0, Centroid = 1, Offset = 2 };
+    Linear = 0,
+    Perspective = 1,
+    Flat = 2,
+    Sc = 3,
+};
+enum class IpaSampleMode : u64 {
+    Default = 0,
+    Centroid = 1,
+    Offset = 2,
+};
 struct IpaMode {
    IpaInterpMode interpolation_mode;
    IpaSampleMode sampling_mode;
-    inline bool operator==(const IpaMode& a) {
-        return (a.interpolation_mode == interpolation_mode) && (a.sampling_mode == sampling_mode);
+    bool operator==(const IpaMode& a) const {
+        return std::tie(interpolation_mode, sampling_mode) ==
+               std::tie(a.interpolation_mode, a.sampling_mode);
    }
-    inline bool operator!=(const IpaMode& a) {
+    bool operator!=(const IpaMode& a) const {
-        return !((*this) == a);
+        return !operator==(a);
    }
 };
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 274c2dbcf..14d82a7bc 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -383,7 +383,7 @@ void RasterizerOpenGL::Clear() {
    bool use_stencil{};
    OpenGLState clear_state;
-    clear_state.draw.draw_framebuffer = state.draw.draw_framebuffer;
+    clear_state.draw.draw_framebuffer = framebuffer.handle;
    clear_state.color_mask.red_enabled = regs.clear_buffers.R ? GL_TRUE : GL_FALSE;
    clear_state.color_mask.green_enabled = regs.clear_buffers.G ? GL_TRUE : GL_FALSE;
    clear_state.color_mask.blue_enabled = regs.clear_buffers.B ? GL_TRUE : GL_FALSE;
@@ -450,6 +450,9 @@ void RasterizerOpenGL::DrawArrays() {
    SyncBlendState();
    SyncLogicOpState();
    SyncCullMode();
+    SyncAlphaTest();
+    SyncTransformFeedback();
+    SyncPointState();
    // TODO(bunnei): Sync framebuffer_scale uniform here
    // TODO(bunnei): Sync scissorbox uniform(s) here
@@ -484,8 +487,13 @@ void RasterizerOpenGL::DrawArrays() {
    GLintptr index_buffer_offset = 0;
    if (is_indexed) {
        MICROPROFILE_SCOPE(OpenGL_Index);
-        index_buffer_offset =
-            buffer_cache.UploadMemory(regs.index_array.StartAddress(), index_buffer_size);
+        // Adjust the index buffer offset so it points to the first desired index.
+        auto index_start = regs.index_array.StartAddress();
+        index_start += static_cast<size_t>(regs.index_array.first) *
+                       static_cast<size_t>(regs.index_array.FormatSizeInBytes());
+        index_buffer_offset = buffer_cache.UploadMemory(index_start, index_buffer_size);
    }
    SetupShaders();
@@ -499,10 +507,6 @@ void RasterizerOpenGL::DrawArrays() {
    if (is_indexed) {
        const GLint base_vertex{static_cast<GLint>(regs.vb_element_base)};
-        // Adjust the index buffer offset so it points to the first desired index.
-        index_buffer_offset += static_cast<GLintptr>(regs.index_array.first) *
-                               static_cast<GLintptr>(regs.index_array.FormatSizeInBytes());
        if (gpu.state.current_instance > 0) {
            glDrawElementsInstancedBaseVertexBaseInstance(
                primitive_mode, regs.index_array.count,
@@ -734,7 +738,7 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader,
        }
        texture_samplers[current_bindpoint].SyncWithConfig(texture.tsc);
-        Surface surface = res_cache.GetTextureSurface(texture);
+        Surface surface = res_cache.GetTextureSurface(texture, entry);
        if (surface != nullptr) {
            state.texture_units[current_bindpoint].texture = surface->Texture().handle;
            state.texture_units[current_bindpoint].target = surface->Target();
@@ -882,4 +886,30 @@ 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::SyncTransformFeedback() {
+    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+    if (regs.tfb_enabled != 0) {
+        LOG_CRITICAL(Render_OpenGL, "Transform feedbacks are not implemented");
+        UNREACHABLE();
+    }
+}
+void RasterizerOpenGL::SyncPointState() {
+    const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
+    state.point.size = regs.point_size;
+}
 } // namespace OpenGL
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h
index bf9560bdc..4c8ecbd1c 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer.h
@@ -158,6 +158,15 @@ 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 transform feedback state to match the guest state
+    void SyncTransformFeedback();
+    /// Syncs the point state to match the guest state
+    void SyncPointState();
    bool has_ARB_direct_state_access = false;
    bool has_ARB_multi_bind = false;
    bool has_ARB_separate_shader_objects = false;
@@ -178,7 +187,7 @@ private:
             OGLVertexArray>
        vertex_array_cache;
-    std::array<SamplerInfo, GLShader::NumTextureSamplers> texture_samplers;
+    std::array<SamplerInfo, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> texture_samplers;
    static constexpr std::size_t STREAM_BUFFER_SIZE = 128 * 1024 * 1024;
    OGLBufferCache buffer_cache;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
index 86682d7cb..ce967c4d6 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
@@ -41,7 +41,7 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
 }
 /*static*/ SurfaceParams SurfaceParams::CreateForTexture(
-    const Tegra::Texture::FullTextureInfo& config) {
+    const Tegra::Texture::FullTextureInfo& config, const GLShader::SamplerEntry& entry) {
    SurfaceParams params{};
    params.addr = TryGetCpuAddr(config.tic.Address());
    params.is_tiled = config.tic.IsTiled();
@@ -60,9 +60,23 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
    case SurfaceTarget::Texture2D:
        params.depth = 1;
        break;
+    case SurfaceTarget::TextureCubemap:
+        params.depth = config.tic.Depth() * 6;
+        break;
    case SurfaceTarget::Texture3D:
+        params.depth = config.tic.Depth();
+        break;
    case SurfaceTarget::Texture2DArray:
        params.depth = config.tic.Depth();
+        if (!entry.IsArray()) {
+            // TODO(bunnei): We have seen games re-use a Texture2D as Texture2DArray with depth of
+            // one, but sample the texture in the shader as if it were not an array texture. This
+            // probably is valid on hardware, but we still need to write a test to confirm this. In
+            // emulation, the workaround here is to continue to treat this as a Texture2D. An
+            // example game that does this is Super Mario Odyssey (in Cloud Kingdom).
+            ASSERT(params.depth == 1);
+            params.target = SurfaceTarget::Texture2D;
+        }
        break;
    default:
        LOG_CRITICAL(HW_GPU, "Unknown depth for target={}", static_cast<u32>(params.target));
@@ -71,7 +85,11 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
        break;
    }
-    params.size_in_bytes = params.SizeInBytes();
+    params.size_in_bytes_total = params.SizeInBytesTotal();
+    params.size_in_bytes_2d = params.SizeInBytes2D();
+    params.max_mip_level = config.tic.max_mip_level + 1;
+    params.rt = {};
    return params;
 }
@@ -89,7 +107,16 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
    params.unaligned_height = config.height;
    params.target = SurfaceTarget::Texture2D;
    params.depth = 1;
-    params.size_in_bytes = params.SizeInBytes();
+    params.size_in_bytes_total = params.SizeInBytesTotal();
+    params.size_in_bytes_2d = params.SizeInBytes2D();
+    params.max_mip_level = 0;
+    // Render target specific parameters, not used for caching
+    params.rt.index = static_cast<u32>(index);
+    params.rt.array_mode = config.array_mode;
+    params.rt.layer_stride = config.layer_stride;
+    params.rt.base_layer = config.base_layer;
    return params;
 }
@@ -108,7 +135,11 @@ static VAddr TryGetCpuAddr(Tegra::GPUVAddr gpu_addr) {
    params.unaligned_height = zeta_height;
    params.target = SurfaceTarget::Texture2D;
    params.depth = 1;
-    params.size_in_bytes = params.SizeInBytes();
+    params.size_in_bytes_total = params.SizeInBytesTotal();
+    params.size_in_bytes_2d = params.SizeInBytes2D();
+    params.max_mip_level = 0;
+    params.rt = {};
    return params;
 }
@@ -141,8 +172,8 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
    {GL_COMPRESSED_RGBA_BPTC_UNORM_ARB, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
     true}, // BC7U
    {GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT_ARB, GL_RGB, GL_UNSIGNED_INT_8_8_8_8,
-     ComponentType::UNorm, true}, // BC6H_UF16
+     ComponentType::Float, true}, // BC6H_UF16
-    {GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT_ARB, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
+    {GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT_ARB, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::Float,
     true},                                                                    // BC6H_SF16
    {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},        // ASTC_2D_4X4
    {GL_RG8, GL_RG, GL_UNSIGNED_BYTE, ComponentType::UNorm, false},            // G8R8U
@@ -400,9 +431,13 @@ static constexpr std::array<void (*)(u32, u32, u32, u8*, std::size_t, VAddr),
        // clang-format on
 };
-static bool BlitTextures(GLuint src_tex, const MathUtil::Rectangle<u32>& src_rect, GLuint dst_tex,
+static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
-                         const MathUtil::Rectangle<u32>& dst_rect, SurfaceType type,
+                        GLuint read_fb_handle, GLuint draw_fb_handle, GLenum src_attachment = 0,
-                         GLuint read_fb_handle, GLuint draw_fb_handle) {
+                        GLenum dst_attachment = 0, std::size_t cubemap_face = 0) {
+    const auto& src_params{src_surface->GetSurfaceParams()};
+    const auto& dst_params{dst_surface->GetSurfaceParams()};
    OpenGLState prev_state{OpenGLState::GetCurState()};
    SCOPE_EXIT({ prev_state.Apply(); });
@@ -413,47 +448,203 @@ static bool BlitTextures(GLuint src_tex, const MathUtil::Rectangle<u32>& src_rec
    u32 buffers{};
-    if (type == SurfaceType::ColorTexture) {
+    if (src_params.type == SurfaceType::ColorTexture) {
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, src_tex,
+        switch (src_params.target) {
-                               0);
+        case SurfaceParams::SurfaceTarget::Texture2D:
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
+            glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
-                               0);
+                                   GL_TEXTURE_2D, src_surface->Texture().handle, 0);
+            glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
+                                   0, 0);
+            break;
+        case SurfaceParams::SurfaceTarget::TextureCubemap:
+            glFramebufferTexture2D(
+                GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
+                static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
+                src_surface->Texture().handle, 0);
+            glFramebufferTexture2D(
+                GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
+                static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
+            break;
+        case SurfaceParams::SurfaceTarget::Texture2DArray:
+            glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
+                                      src_surface->Texture().handle, 0, 0);
+            glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
+            break;
+        case SurfaceParams::SurfaceTarget::Texture3D:
+            glFramebufferTexture3D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
+                                   SurfaceTargetToGL(src_params.target),
+                                   src_surface->Texture().handle, 0, 0);
+            glFramebufferTexture3D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
+                                   SurfaceTargetToGL(src_params.target), 0, 0, 0);
+            break;
+        default:
+            glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
+                                   GL_TEXTURE_2D, src_surface->Texture().handle, 0);
+            glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
+                                   0, 0);
+            break;
+        }
+        switch (dst_params.target) {
+        case SurfaceParams::SurfaceTarget::Texture2D:
+            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
+                                   GL_TEXTURE_2D, dst_surface->Texture().handle, 0);
+            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
+                                   0, 0);
+            break;
+        case SurfaceParams::SurfaceTarget::TextureCubemap:
+            glFramebufferTexture2D(
+                GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
+                static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
+                dst_surface->Texture().handle, 0);
+            glFramebufferTexture2D(
+                GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
+                static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
+            break;
+        case SurfaceParams::SurfaceTarget::Texture2DArray:
+            glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
+                                      dst_surface->Texture().handle, 0, 0);
+            glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
+            break;
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, dst_tex,
+        case SurfaceParams::SurfaceTarget::Texture3D:
-                               0);
+            glFramebufferTexture3D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0,
+                                   SurfaceTargetToGL(dst_params.target),
-                               0);
+                                   dst_surface->Texture().handle, 0, 0);
+            glFramebufferTexture3D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
+                                   SurfaceTargetToGL(dst_params.target), 0, 0, 0);
+            break;
+        default:
+            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
+                                   GL_TEXTURE_2D, dst_surface->Texture().handle, 0);
+            glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
+                                   0, 0);
+            break;
+        }
        buffers = GL_COLOR_BUFFER_BIT;
-    } else if (type == SurfaceType::Depth) {
+    } else if (src_params.type == SurfaceType::Depth) {
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
+        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D, src_tex, 0);
+                               GL_TEXTURE_2D, 0, 0);
+        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D,
+                               src_surface->Texture().handle, 0);
        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0);
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
+        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D, dst_tex, 0);
+                               GL_TEXTURE_2D, 0, 0);
+        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D,
+                               dst_surface->Texture().handle, 0);
        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0);
        buffers = GL_DEPTH_BUFFER_BIT;
-    } else if (type == SurfaceType::DepthStencil) {
+    } else if (src_params.type == SurfaceType::DepthStencil) {
-        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
+        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
+                               GL_TEXTURE_2D, 0, 0);
        glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
-                               src_tex, 0);
+                               src_surface->Texture().handle, 0);
-        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0);
+        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
+                               GL_TEXTURE_2D, 0, 0);
        glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
-                               dst_tex, 0);
+                               dst_surface->Texture().handle, 0);
        buffers = GL_DEPTH_BUFFER_BIT | GL_STENCIL_BUFFER_BIT;
    }
-    glBlitFramebuffer(src_rect.left, src_rect.bottom, src_rect.right, src_rect.top, dst_rect.left,
+    const auto& rect{src_params.GetRect()};
-                      dst_rect.bottom, dst_rect.right, dst_rect.top, buffers,
+    glBlitFramebuffer(rect.left, rect.bottom, rect.right, rect.top, rect.left, rect.bottom,
+                      rect.right, rect.top, buffers,
                      buffers == GL_COLOR_BUFFER_BIT ? GL_LINEAR : GL_NEAREST);
    return true;
 }
+static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
+                        GLuint copy_pbo_handle, GLenum src_attachment = 0,
+                        GLenum dst_attachment = 0, std::size_t cubemap_face = 0) {
+    ASSERT_MSG(dst_attachment == 0, "Unimplemented");
+    const auto& src_params{src_surface->GetSurfaceParams()};
+    const auto& dst_params{dst_surface->GetSurfaceParams()};
+    auto source_format = GetFormatTuple(src_params.pixel_format, src_params.component_type);
+    auto dest_format = GetFormatTuple(dst_params.pixel_format, dst_params.component_type);
+    std::size_t buffer_size =
+        std::max(src_params.size_in_bytes_total, dst_params.size_in_bytes_total);
+    glBindBuffer(GL_PIXEL_PACK_BUFFER, copy_pbo_handle);
+    glBufferData(GL_PIXEL_PACK_BUFFER, buffer_size, nullptr, GL_STREAM_DRAW_ARB);
+    if (source_format.compressed) {
+        glGetCompressedTextureImage(src_surface->Texture().handle, src_attachment,
+                                    static_cast<GLsizei>(src_params.size_in_bytes_total), nullptr);
+    } else {
+        glGetTextureImage(src_surface->Texture().handle, src_attachment, source_format.format,
+                          source_format.type, static_cast<GLsizei>(src_params.size_in_bytes_total),
+                          nullptr);
+    }
+    // If the new texture is bigger than the previous one, we need to fill in the rest with data
+    // from the CPU.
+    if (src_params.size_in_bytes_total < dst_params.size_in_bytes_total) {
+        // Upload the rest of the memory.
+        if (dst_params.is_tiled) {
+            // TODO(Subv): We might have to de-tile the subtexture and re-tile it with the rest
+            // of the data in this case. Games like Super Mario Odyssey seem to hit this case
+            // when drawing, it re-uses the memory of a previous texture as a bigger framebuffer
+            // but it doesn't clear it beforehand, the texture is already full of zeros.
+            LOG_DEBUG(HW_GPU, "Trying to upload extra texture data from the CPU during "
+                              "reinterpretation but the texture is tiled.");
+        }
+        std::size_t remaining_size =
+            dst_params.size_in_bytes_total - src_params.size_in_bytes_total;
+        std::vector<u8> data(remaining_size);
+        Memory::ReadBlock(dst_params.addr + src_params.size_in_bytes_total, data.data(),
+                          data.size());
+        glBufferSubData(GL_PIXEL_PACK_BUFFER, src_params.size_in_bytes_total, remaining_size,
+                        data.data());
+    }
+    glBindBuffer(GL_PIXEL_PACK_BUFFER, 0);
+    const GLsizei width{static_cast<GLsizei>(
+        std::min(src_params.GetRect().GetWidth(), dst_params.GetRect().GetWidth()))};
+    const GLsizei height{static_cast<GLsizei>(
+        std::min(src_params.GetRect().GetHeight(), dst_params.GetRect().GetHeight()))};
+    glBindBuffer(GL_PIXEL_UNPACK_BUFFER, copy_pbo_handle);
+    if (dest_format.compressed) {
+        LOG_CRITICAL(HW_GPU, "Compressed copy is unimplemented!");
+        UNREACHABLE();
+    } else {
+        switch (dst_params.target) {
+        case SurfaceParams::SurfaceTarget::Texture1D:
+            glTextureSubImage1D(dst_surface->Texture().handle, 0, 0, width, dest_format.format,
+                                dest_format.type, nullptr);
+            break;
+        case SurfaceParams::SurfaceTarget::Texture2D:
+            glTextureSubImage2D(dst_surface->Texture().handle, 0, 0, 0, width, height,
+                                dest_format.format, dest_format.type, nullptr);
+            break;
+        case SurfaceParams::SurfaceTarget::Texture3D:
+        case SurfaceParams::SurfaceTarget::Texture2DArray:
+            glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0, 0, width, height,
+                                static_cast<GLsizei>(dst_params.depth), dest_format.format,
+                                dest_format.type, nullptr);
+            break;
+        case SurfaceParams::SurfaceTarget::TextureCubemap:
+            glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0,
+                                static_cast<GLint>(cubemap_face), width, height, 1,
+                                dest_format.format, dest_format.type, nullptr);
+            break;
+        default:
+            LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
+                         static_cast<u32>(dst_params.target));
+            UNREACHABLE();
+        }
+        glBindBuffer(GL_PIXEL_UNPACK_BUFFER, 0);
+    }
+}
 CachedSurface::CachedSurface(const SurfaceParams& params)
    : params(params), gl_target(SurfaceTargetToGL(params.target)) {
    texture.Create();
@@ -481,6 +672,7 @@ CachedSurface::CachedSurface(const SurfaceParams& params)
                           rect.GetWidth());
            break;
        case SurfaceParams::SurfaceTarget::Texture2D:
+        case SurfaceParams::SurfaceTarget::TextureCubemap:
            glTexStorage2D(SurfaceTargetToGL(params.target), 1, format_tuple.internal_format,
                           rect.GetWidth(), rect.GetHeight());
            break;
@@ -501,6 +693,9 @@ CachedSurface::CachedSurface(const SurfaceParams& params)
    glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_MIN_FILTER, GL_LINEAR);
    glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
    glTexParameteri(SurfaceTargetToGL(params.target), GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    VideoCore::LabelGLObject(GL_TEXTURE, texture.handle, params.addr,
+                             SurfaceParams::SurfaceTargetName(params.target));
 }
 static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height) {
@@ -582,29 +777,39 @@ void CachedSurface::LoadGLBuffer() {
    const u32 bytes_per_pixel = GetGLBytesPerPixel(params.pixel_format);
    const u32 copy_size = params.width * params.height * bytes_per_pixel;
+    const std::size_t total_size = copy_size * params.depth;
    MICROPROFILE_SCOPE(OpenGL_SurfaceLoad);
    if (params.is_tiled) {
+        gl_buffer.resize(total_size);
        // TODO(bunnei): This only unswizzles and copies a 2D texture - we do not yet know how to do
        // this for 3D textures, etc.
        switch (params.target) {
        case SurfaceParams::SurfaceTarget::Texture2D:
            // Pass impl. to the fallback code below
            break;
+        case SurfaceParams::SurfaceTarget::Texture2DArray:
+        case SurfaceParams::SurfaceTarget::TextureCubemap:
+            for (std::size_t index = 0; index < params.depth; ++index) {
+                const std::size_t offset{index * copy_size};
+                morton_to_gl_fns[static_cast<std::size_t>(params.pixel_format)](
+                    params.width, params.block_height, params.height, gl_buffer.data() + offset,
+                    copy_size, params.addr + offset);
+            }
+            break;
        default:
            LOG_CRITICAL(HW_GPU, "Unimplemented tiled load for target={}",
                         static_cast<u32>(params.target));
            UNREACHABLE();
        }
-        gl_buffer.resize(static_cast<std::size_t>(params.depth) * copy_size);
        morton_to_gl_fns[static_cast<std::size_t>(params.pixel_format)](
            params.width, params.block_height, params.height, gl_buffer.data(), copy_size,
            params.addr);
    } else {
-        const u8* const texture_src_data_end{texture_src_data +
+        const u8* const texture_src_data_end{texture_src_data + total_size};
-                                             (static_cast<std::size_t>(params.depth) * copy_size)};
        gl_buffer.assign(texture_src_data, texture_src_data_end);
    }
@@ -631,7 +836,7 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle
    // Load data from memory to the surface
    const GLint x0 = static_cast<GLint>(rect.left);
    const GLint y0 = static_cast<GLint>(rect.bottom);
-    const std::size_t buffer_offset =
+    std::size_t buffer_offset =
        static_cast<std::size_t>(static_cast<std::size_t>(y0) * params.width +
                                 static_cast<std::size_t>(x0)) *
        GetGLBytesPerPixel(params.pixel_format);
@@ -660,15 +865,25 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle
            glCompressedTexImage2D(
                SurfaceTargetToGL(params.target), 0, tuple.internal_format,
                static_cast<GLsizei>(params.width), static_cast<GLsizei>(params.height), 0,
-                static_cast<GLsizei>(params.size_in_bytes), &gl_buffer[buffer_offset]);
+                static_cast<GLsizei>(params.size_in_bytes_2d), &gl_buffer[buffer_offset]);
            break;
        case SurfaceParams::SurfaceTarget::Texture3D:
        case SurfaceParams::SurfaceTarget::Texture2DArray:
            glCompressedTexImage3D(
                SurfaceTargetToGL(params.target), 0, tuple.internal_format,
                static_cast<GLsizei>(params.width), static_cast<GLsizei>(params.height),
-                static_cast<GLsizei>(params.depth), 0, static_cast<GLsizei>(params.size_in_bytes),
+                static_cast<GLsizei>(params.depth), 0,
-                &gl_buffer[buffer_offset]);
+                static_cast<GLsizei>(params.size_in_bytes_total), &gl_buffer[buffer_offset]);
+            break;
+        case SurfaceParams::SurfaceTarget::TextureCubemap:
+            for (std::size_t face = 0; face < params.depth; ++face) {
+                glCompressedTexImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face),
+                                       0, tuple.internal_format, static_cast<GLsizei>(params.width),
+                                       static_cast<GLsizei>(params.height), 0,
+                                       static_cast<GLsizei>(params.size_in_bytes_2d),
+                                       &gl_buffer[buffer_offset]);
+                buffer_offset += params.size_in_bytes_2d;
+            }
            break;
        default:
            LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
@@ -676,8 +891,8 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle
            UNREACHABLE();
            glCompressedTexImage2D(
                GL_TEXTURE_2D, 0, tuple.internal_format, static_cast<GLsizei>(params.width),
-                static_cast<GLsizei>(params.height), 0, static_cast<GLsizei>(params.size_in_bytes),
+                static_cast<GLsizei>(params.height), 0,
-                &gl_buffer[buffer_offset]);
+                static_cast<GLsizei>(params.size_in_bytes_2d), &gl_buffer[buffer_offset]);
        }
    } else {
@@ -700,6 +915,15 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle
                            static_cast<GLsizei>(rect.GetHeight()), params.depth, tuple.format,
                            tuple.type, &gl_buffer[buffer_offset]);
            break;
+        case SurfaceParams::SurfaceTarget::TextureCubemap:
+            for (std::size_t face = 0; face < params.depth; ++face) {
+                glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), 0, x0,
+                                y0, static_cast<GLsizei>(rect.GetWidth()),
+                                static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
+                                &gl_buffer[buffer_offset]);
+                buffer_offset += params.size_in_bytes_2d;
+            }
+            break;
        default:
            LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
                         static_cast<u32>(params.target));
@@ -719,8 +943,9 @@ RasterizerCacheOpenGL::RasterizerCacheOpenGL() {
    copy_pbo.Create();
 }
-Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextureInfo& config) {
+Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextureInfo& config,
-    return GetSurface(SurfaceParams::CreateForTexture(config));
+                                                 const GLShader::SamplerEntry& entry) {
+    return GetSurface(SurfaceParams::CreateForTexture(config, entry));
 }
 Surface RasterizerCacheOpenGL::GetDepthBufferSurface(bool preserve_contents) {
@@ -808,98 +1033,69 @@ Surface RasterizerCacheOpenGL::GetUncachedSurface(const SurfaceParams& params) {
    return surface;
 }
-Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& surface,
+Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
                                               const SurfaceParams& new_params) {
    // Verify surface is compatible for blitting
-    const auto& params{surface->GetSurfaceParams()};
+    auto old_params{old_surface->GetSurfaceParams()};
    // Get a new surface with the new parameters, and blit the previous surface to it
    Surface new_surface{GetUncachedSurface(new_params)};
-    if (params.pixel_format == new_params.pixel_format ||
+    // If the format is the same, just do a framebuffer blit. This is significantly faster than
-        !Settings::values.use_accurate_framebuffers) {
+    // using PBOs. The is also likely less accurate, as textures will be converted rather than
-        // If the format is the same, just do a framebuffer blit. This is significantly faster than
+    // reinterpreted. When use_accurate_framebuffers setting is enabled, perform a more accurate
-        // using PBOs. The is also likely less accurate, as textures will be converted rather than
+    // surface copy, where pixels are reinterpreted as a new format (without conversion). This
-        // reinterpreted.
+    // code path uses OpenGL PBOs and is quite slow.
+    const bool is_blit{old_params.pixel_format == new_params.pixel_format ||
-        BlitTextures(surface->Texture().handle, params.GetRect(), new_surface->Texture().handle,
+                       !Settings::values.use_accurate_framebuffers};
-                     params.GetRect(), params.type, read_framebuffer.handle,
-                     draw_framebuffer.handle);
-    } else {
-        // When use_accurate_framebuffers setting is enabled, perform a more accurate surface copy,
-        // where pixels are reinterpreted as a new format (without conversion). This code path uses
-        // OpenGL PBOs and is quite slow.
-        auto source_format = GetFormatTuple(params.pixel_format, params.component_type);
-        auto dest_format = GetFormatTuple(new_params.pixel_format, new_params.component_type);
-        std::size_t buffer_size = std::max(params.SizeInBytes(), new_params.SizeInBytes());
+    switch (new_params.target) {
+    case SurfaceParams::SurfaceTarget::Texture2D:
-        glBindBuffer(GL_PIXEL_PACK_BUFFER, copy_pbo.handle);
+        if (is_blit) {
-        glBufferData(GL_PIXEL_PACK_BUFFER, buffer_size, nullptr, GL_STREAM_DRAW_ARB);
+            BlitSurface(old_surface, new_surface, read_framebuffer.handle, draw_framebuffer.handle);
-        if (source_format.compressed) {
-            glGetCompressedTextureImage(surface->Texture().handle, 0,
-                                        static_cast<GLsizei>(params.SizeInBytes()), nullptr);
        } else {
-            glGetTextureImage(surface->Texture().handle, 0, source_format.format,
+            CopySurface(old_surface, new_surface, copy_pbo.handle);
-                              source_format.type, static_cast<GLsizei>(params.SizeInBytes()),
-                              nullptr);
        }
-        // If the new texture is bigger than the previous one, we need to fill in the rest with data
+        break;
-        // from the CPU.
+    case SurfaceParams::SurfaceTarget::TextureCubemap: {
-        if (params.SizeInBytes() < new_params.SizeInBytes()) {
+        if (old_params.rt.array_mode != 1) {
-            // Upload the rest of the memory.
+            // TODO(bunnei): This is used by Breath of the Wild, I'm not sure how to implement this
-            if (new_params.is_tiled) {
+            // yet (array rendering used as a cubemap texture).
-                // TODO(Subv): We might have to de-tile the subtexture and re-tile it with the rest
+            LOG_CRITICAL(HW_GPU, "Unhandled rendertarget array_mode {}", old_params.rt.array_mode);
-                // of the data in this case. Games like Super Mario Odyssey seem to hit this case
+            UNREACHABLE();
-                // when drawing, it re-uses the memory of a previous texture as a bigger framebuffer
+            return new_surface;
-                // but it doesn't clear it beforehand, the texture is already full of zeros.
-                LOG_DEBUG(HW_GPU, "Trying to upload extra texture data from the CPU during "
-                                  "reinterpretation but the texture is tiled.");
-            }
-            std::size_t remaining_size = new_params.SizeInBytes() - params.SizeInBytes();
-            std::vector<u8> data(remaining_size);
-            Memory::ReadBlock(new_params.addr + params.SizeInBytes(), data.data(), data.size());
-            glBufferSubData(GL_PIXEL_PACK_BUFFER, params.SizeInBytes(), remaining_size,
-                            data.data());
        }
-        glBindBuffer(GL_PIXEL_PACK_BUFFER, 0);
+        // This seems to be used for render-to-cubemap texture
+        ASSERT_MSG(old_params.target == SurfaceParams::SurfaceTarget::Texture2D, "Unexpected");
-        const auto& dest_rect{new_params.GetRect()};
+        ASSERT_MSG(old_params.pixel_format == new_params.pixel_format, "Unexpected");
+        ASSERT_MSG(old_params.rt.base_layer == 0, "Unimplemented");
-        glBindBuffer(GL_PIXEL_UNPACK_BUFFER, copy_pbo.handle);
-        if (dest_format.compressed) {
+        // TODO(bunnei): Verify the below - this stride seems to be in 32-bit words, not pixels.
-            LOG_CRITICAL(HW_GPU, "Compressed copy is unimplemented!");
+        // Tested with Splatoon 2, Super Mario Odyssey, and Breath of the Wild.
-            UNREACHABLE();
+        const std::size_t byte_stride{old_params.rt.layer_stride * sizeof(u32)};
-        } else {
-            switch (new_params.target) {
+        for (std::size_t index = 0; index < new_params.depth; ++index) {
-            case SurfaceParams::SurfaceTarget::Texture1D:
+            Surface face_surface{TryGetReservedSurface(old_params)};
-                glTextureSubImage1D(new_surface->Texture().handle, 0, 0,
+            ASSERT_MSG(face_surface, "Unexpected");
-                                    static_cast<GLsizei>(dest_rect.GetWidth()), dest_format.format,
-                                    dest_format.type, nullptr);
+            if (is_blit) {
-                break;
+                BlitSurface(face_surface, new_surface, read_framebuffer.handle,
-            case SurfaceParams::SurfaceTarget::Texture2D:
+                            draw_framebuffer.handle, face_surface->GetSurfaceParams().rt.index,
-                glTextureSubImage2D(new_surface->Texture().handle, 0, 0, 0,
+                            new_params.rt.index, index);
-                                    static_cast<GLsizei>(dest_rect.GetWidth()),
+            } else {
-                                    static_cast<GLsizei>(dest_rect.GetHeight()), dest_format.format,
+                CopySurface(face_surface, new_surface, copy_pbo.handle,
-                                    dest_format.type, nullptr);
+                            face_surface->GetSurfaceParams().rt.index, new_params.rt.index, index);
-                break;
-            case SurfaceParams::SurfaceTarget::Texture3D:
-            case SurfaceParams::SurfaceTarget::Texture2DArray:
-                glTextureSubImage3D(new_surface->Texture().handle, 0, 0, 0, 0,
-                                    static_cast<GLsizei>(dest_rect.GetWidth()),
-                                    static_cast<GLsizei>(dest_rect.GetHeight()),
-                                    static_cast<GLsizei>(new_params.depth), dest_format.format,
-                                    dest_format.type, nullptr);
-                break;
-            default:
-                LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
-                             static_cast<u32>(params.target));
-                UNREACHABLE();
            }
+            old_params.addr += byte_stride;
        }
-        glBindBuffer(GL_PIXEL_UNPACK_BUFFER, 0);
+        break;
+    }
+    default:
+        LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
+                     static_cast<u32>(new_params.target));
+        UNREACHABLE();
    }
    return new_surface;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
index d7a4bc37f..49025a3fe 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
@@ -9,12 +9,14 @@
 #include <memory>
 #include <vector>
+#include "common/alignment.h"
 #include "common/common_types.h"
 #include "common/hash.h"
 #include "common/math_util.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/rasterizer_cache.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
+#include "video_core/renderer_opengl/gl_shader_gen.h"
 #include "video_core/textures/texture.h"
 namespace OpenGL {
@@ -126,6 +128,8 @@ struct SurfaceParams {
        case Tegra::Texture::TextureType::Texture2D:
        case Tegra::Texture::TextureType::Texture2DNoMipmap:
            return SurfaceTarget::Texture2D;
+        case Tegra::Texture::TextureType::TextureCubemap:
+            return SurfaceTarget::TextureCubemap;
        case Tegra::Texture::TextureType::Texture1DArray:
            return SurfaceTarget::Texture1DArray;
        case Tegra::Texture::TextureType::Texture2DArray:
@@ -137,6 +141,27 @@ struct SurfaceParams {
        }
    }
+    static std::string SurfaceTargetName(SurfaceTarget target) {
+        switch (target) {
+        case SurfaceTarget::Texture1D:
+            return "Texture1D";
+        case SurfaceTarget::Texture2D:
+            return "Texture2D";
+        case SurfaceTarget::Texture3D:
+            return "Texture3D";
+        case SurfaceTarget::Texture1DArray:
+            return "Texture1DArray";
+        case SurfaceTarget::Texture2DArray:
+            return "Texture2DArray";
+        case SurfaceTarget::TextureCubemap:
+            return "TextureCubemap";
+        default:
+            LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
+            UNREACHABLE();
+            return fmt::format("TextureUnknown({})", static_cast<u32>(target));
+        }
+    }
    /**
     * 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
@@ -668,17 +693,23 @@ struct SurfaceParams {
    /// Returns the rectangle corresponding to this surface
    MathUtil::Rectangle<u32> GetRect() const;
-    /// Returns the size of this surface in bytes, adjusted for compression
+    /// Returns the size of this surface as a 2D texture in bytes, adjusted for compression
-    std::size_t SizeInBytes() const {
+    std::size_t SizeInBytes2D() const {
        const u32 compression_factor{GetCompressionFactor(pixel_format)};
        ASSERT(width % compression_factor == 0);
        ASSERT(height % compression_factor == 0);
        return (width / compression_factor) * (height / compression_factor) *
-               GetFormatBpp(pixel_format) * depth / CHAR_BIT;
+               GetFormatBpp(pixel_format) / CHAR_BIT;
+    }
+    /// Returns the total size of this surface in bytes, adjusted for compression
+    std::size_t SizeInBytesTotal() const {
+        return SizeInBytes2D() * depth;
    }
    /// Creates SurfaceParams from a texture configuration
-    static SurfaceParams CreateForTexture(const Tegra::Texture::FullTextureInfo& config);
+    static SurfaceParams CreateForTexture(const Tegra::Texture::FullTextureInfo& config,
+                                          const GLShader::SamplerEntry& entry);
    /// Creates SurfaceParams from a framebuffer configuration
    static SurfaceParams CreateForFramebuffer(std::size_t index);
@@ -690,8 +721,9 @@ struct SurfaceParams {
    /// Checks if surfaces are compatible for caching
    bool IsCompatibleSurface(const SurfaceParams& other) const {
-        return std::tie(pixel_format, type, width, height) ==
+        return std::tie(pixel_format, type, width, height, target, depth) ==
-               std::tie(other.pixel_format, other.type, other.width, other.height);
+               std::tie(other.pixel_format, other.type, other.width, other.height, other.target,
+                        other.depth);
    }
    VAddr addr;
@@ -704,8 +736,18 @@ struct SurfaceParams {
    u32 height;
    u32 depth;
    u32 unaligned_height;
-    std::size_t size_in_bytes;
+    std::size_t size_in_bytes_total;
+    std::size_t size_in_bytes_2d;
    SurfaceTarget target;
+    u32 max_mip_level;
+    // Render target specific parameters, not used in caching
+    struct {
+        u32 index;
+        u32 array_mode;
+        u32 layer_stride;
+        u32 base_layer;
+    } rt;
 };
 }; // namespace OpenGL
@@ -715,6 +757,7 @@ struct SurfaceReserveKey : Common::HashableStruct<OpenGL::SurfaceParams> {
    static SurfaceReserveKey Create(const OpenGL::SurfaceParams& params) {
        SurfaceReserveKey res;
        res.state = params;
+        res.state.rt = {}; // Ignore rt config in caching
        return res;
    }
 };
@@ -738,7 +781,7 @@ public:
    }
    std::size_t GetSizeInBytes() const {
-        return params.size_in_bytes;
+        return params.size_in_bytes_total;
    }
    const OGLTexture& Texture() const {
@@ -779,7 +822,8 @@ public:
    RasterizerCacheOpenGL();
    /// Get a surface based on the texture configuration
-    Surface GetTextureSurface(const Tegra::Texture::FullTextureInfo& config);
+    Surface GetTextureSurface(const Tegra::Texture::FullTextureInfo& config,
+                              const GLShader::SamplerEntry& entry);
    /// Get the depth surface based on the framebuffer configuration
    Surface GetDepthBufferSurface(bool preserve_contents);
@@ -801,7 +845,7 @@ private:
    Surface GetUncachedSurface(const SurfaceParams& params);
    /// Recreates a surface with new parameters
-    Surface RecreateSurface(const Surface& surface, const SurfaceParams& new_params);
+    Surface RecreateSurface(const Surface& old_surface, const SurfaceParams& new_params);
    /// Reserves a unique surface that can be reused later
    void ReserveSurface(const Surface& surface);
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index 894fe6eae..7cd8f91e4 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -8,6 +8,7 @@
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_opengl/gl_shader_cache.h"
 #include "video_core/renderer_opengl/gl_shader_manager.h"
+#include "video_core/utils.h"
 namespace OpenGL {
@@ -83,6 +84,7 @@ CachedShader::CachedShader(VAddr addr, Maxwell::ShaderProgram program_type)
    shader.Create(program_result.first.c_str(), gl_type);
    program.Create(true, shader.handle);
    SetShaderUniformBlockBindings(program.handle);
+    VideoCore::LabelGLObject(GL_PROGRAM, program.handle, addr);
 }
 GLuint CachedShader::GetProgramResourceIndex(const GLShader::ConstBufferEntry& buffer) {
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
index 702ffbe9c..579a78702 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
@@ -2000,6 +2000,14 @@ private:
                    }
                    break;
                }
+                case Tegra::Shader::TextureType::TextureCube: {
+                    ASSERT_MSG(!is_array, "Unimplemented");
+                    std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+                    std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+                    std::string z = regs.GetRegisterAsFloat(instr.gpr20);
+                    coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+                    break;
+                }
                default:
                    LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
                                 static_cast<u32>(texture_type));
diff --git a/src/video_core/renderer_opengl/gl_shader_manager.h b/src/video_core/renderer_opengl/gl_shader_manager.h
index b86cd96e8..3de15ba9b 100644
--- a/src/video_core/renderer_opengl/gl_shader_manager.h
+++ b/src/video_core/renderer_opengl/gl_shader_manager.h
@@ -11,9 +11,6 @@
 namespace OpenGL::GLShader {
-/// Number of OpenGL texture samplers that can be used in the fragment shader
-static constexpr std::size_t NumTextureSamplers = 32;
 using Tegra::Engines::Maxwell3D;
 /// Uniform structure for the Uniform Buffer Object, all vectors must be 16-byte aligned
diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp
index af99132ba..1fe26a2a9 100644
--- a/src/video_core/renderer_opengl/gl_state.cpp
+++ b/src/video_core/renderer_opengl/gl_state.cpp
@@ -4,6 +4,7 @@
 #include <iterator>
 #include <glad/glad.h>
+#include "common/assert.h"
 #include "common/logging/log.h"
 #include "video_core/renderer_opengl/gl_state.h"
@@ -78,6 +79,8 @@ OpenGLState::OpenGLState() {
    viewport.height = 0;
    clip_distance = {};
+    point.size = 1;
 }
 void OpenGLState::Apply() const {
@@ -204,9 +207,6 @@ void OpenGLState::Apply() const {
            glActiveTexture(TextureUnits::MaxwellTexture(static_cast<int>(i)).Enum());
            glBindTexture(texture_unit.target, texture_unit.texture);
        }
-        if (texture_unit.sampler != cur_state_texture_unit.sampler) {
-            glBindSampler(static_cast<GLuint>(i), texture_unit.sampler);
-        }
        // Update the texture swizzle
        if (texture_unit.swizzle.r != cur_state_texture_unit.swizzle.r ||
            texture_unit.swizzle.g != cur_state_texture_unit.swizzle.g ||
@@ -218,6 +218,27 @@ void OpenGLState::Apply() const {
        }
    }
+    // Samplers
+    {
+        bool has_delta{};
+        std::size_t first{}, last{};
+        std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> samplers;
+        for (std::size_t i = 0; i < std::size(samplers); ++i) {
+            samplers[i] = texture_units[i].sampler;
+            if (samplers[i] != cur_state.texture_units[i].sampler) {
+                if (!has_delta) {
+                    first = i;
+                    has_delta = true;
+                }
+                last = i;
+            }
+        }
+        if (has_delta) {
+            glBindSamplers(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1),
+                           samplers.data());
+        }
+    }
    // Framebuffer
    if (draw.read_framebuffer != cur_state.draw.read_framebuffer) {
        glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer);
@@ -282,6 +303,11 @@ void OpenGLState::Apply() const {
        }
    }
+    // Point
+    if (point.size != cur_state.point.size) {
+        glPointSize(point.size);
+    }
    cur_state = *this;
 }
diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h
index e3e24b9e7..dc21a2ee3 100644
--- a/src/video_core/renderer_opengl/gl_state.h
+++ b/src/video_core/renderer_opengl/gl_state.h
@@ -6,13 +6,10 @@
 #include <array>
 #include <glad/glad.h>
 #include "video_core/engines/maxwell_3d.h"
 namespace OpenGL {
-using Regs = Tegra::Engines::Maxwell3D::Regs;
 namespace TextureUnits {
 struct TextureUnit {
@@ -118,7 +115,7 @@ public:
            target = GL_TEXTURE_2D;
        }
    };
-    std::array<TextureUnit, 32> texture_units;
+    std::array<TextureUnit, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> texture_units;
    struct {
        GLuint read_framebuffer; // GL_READ_FRAMEBUFFER_BINDING
@@ -145,6 +142,10 @@ public:
        GLsizei height;
    } viewport;
+    struct {
+        float size; // GL_POINT_SIZE
+    } point;
    std::array<bool, 2> clip_distance; // GL_CLIP_DISTANCE
    OpenGLState();
diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.cpp b/src/video_core/renderer_opengl/gl_stream_buffer.cpp
index 664f3ca20..e409228cc 100644
--- a/src/video_core/renderer_opengl/gl_stream_buffer.cpp
+++ b/src/video_core/renderer_opengl/gl_stream_buffer.cpp
@@ -74,7 +74,7 @@ std::tuple<u8*, GLintptr, bool> OGLStreamBuffer::Map(GLsizeiptr size, GLintptr a
        }
    }
-    if (invalidate | !persistent) {
+    if (invalidate || !persistent) {
        GLbitfield flags = GL_MAP_WRITE_BIT | (persistent ? GL_MAP_PERSISTENT_BIT : 0) |
                           (coherent ? GL_MAP_COHERENT_BIT : GL_MAP_FLUSH_EXPLICIT_BIT) |
                           (invalidate ? GL_MAP_INVALIDATE_BUFFER_BIT : GL_MAP_UNSYNCHRONIZED_BIT);
diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp
index 20ba6d4f6..3d5476e5d 100644
--- a/src/video_core/textures/decoders.cpp
+++ b/src/video_core/textures/decoders.cpp
@@ -13,47 +13,20 @@
 namespace Tegra::Texture {
 /**
+ * This table represents the internal swizzle of a gob,
+ * in format 16 bytes x 2 sector packing.
 * Calculates the offset of an (x, y) position within a swizzled texture.
- * Taken from the Tegra X1 TRM.
+ * Taken from the Tegra X1 Technical Reference Manual. pages 1187-1188
 */
-static u32 GetSwizzleOffset(u32 x, u32 y, u32 image_width, u32 bytes_per_pixel, u32 block_height) {
+template <std::size_t N, std::size_t M, u32 Align>
-    // Round up to the next gob
-    const u32 image_width_in_gobs{(image_width * bytes_per_pixel + 63) / 64};
-    u32 GOB_address = 0 + (y / (8 * block_height)) * 512 * block_height * image_width_in_gobs +
-                      (x * bytes_per_pixel / 64) * 512 * block_height +
-                      (y % (8 * block_height) / 8) * 512;
-    x *= bytes_per_pixel;
-    u32 address = GOB_address + ((x % 64) / 32) * 256 + ((y % 8) / 2) * 64 + ((x % 32) / 16) * 32 +
-                  (y % 2) * 16 + (x % 16);
-    return address;
-}
-void CopySwizzledData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                      u8* swizzled_data, u8* unswizzled_data, bool unswizzle, u32 block_height) {
-    u8* data_ptrs[2];
-    for (unsigned y = 0; y < height; ++y) {
-        for (unsigned x = 0; x < width; ++x) {
-            u32 swizzle_offset = GetSwizzleOffset(x, y, width, bytes_per_pixel, block_height);
-            u32 pixel_index = (x + y * width) * out_bytes_per_pixel;
-            data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
-            data_ptrs[!unswizzle] = &unswizzled_data[pixel_index];
-            std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
-        }
-    }
-}
-template <std::size_t N, std::size_t M>
 struct alignas(64) SwizzleTable {
+    static_assert(M * Align == 64, "Swizzle Table does not align to GOB");
    constexpr SwizzleTable() {
        for (u32 y = 0; y < N; ++y) {
            for (u32 x = 0; x < M; ++x) {
-                const u32 x2 = x * 16;
+                const u32 x2 = x * Align;
                values[y][x] = static_cast<u16>(((x2 % 64) / 32) * 256 + ((y % 8) / 2) * 64 +
-                                                ((x2 % 32) / 16) * 32 + (y % 2) * 16);
+                                                ((x2 % 32) / 16) * 32 + (y % 2) * 16 + (x2 % 16));
            }
        }
    }
@@ -63,24 +36,60 @@ struct alignas(64) SwizzleTable {
    std::array<std::array<u16, M>, N> values{};
 };
-constexpr auto swizzle_table = SwizzleTable<8, 4>();
+constexpr auto legacy_swizzle_table = SwizzleTable<8, 64, 1>();
+constexpr auto fast_swizzle_table = SwizzleTable<8, 4, 16>();
-void FastSwizzleData(u32 width, u32 height, u32 bytes_per_pixel, u8* swizzled_data,
+static void LegacySwizzleData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
-                     u8* unswizzled_data, bool unswizzle, u32 block_height) {
+                              u8* swizzled_data, u8* unswizzled_data, bool unswizzle,
+                              u32 block_height) {
+    std::array<u8*, 2> data_ptrs;
+    const std::size_t stride = width * bytes_per_pixel;
+    const std::size_t gobs_in_x = 64;
+    const std::size_t gobs_in_y = 8;
+    const std::size_t gobs_size = gobs_in_x * gobs_in_y;
+    const std::size_t image_width_in_gobs{(stride + gobs_in_x - 1) / gobs_in_x};
+    for (std::size_t y = 0; y < height; ++y) {
+        const std::size_t gob_y_address =
+            (y / (gobs_in_y * block_height)) * gobs_size * block_height * image_width_in_gobs +
+            (y % (gobs_in_y * block_height) / gobs_in_y) * gobs_size;
+        const auto& table = legacy_swizzle_table[y % gobs_in_y];
+        for (std::size_t x = 0; x < width; ++x) {
+            const std::size_t gob_address =
+                gob_y_address + (x * bytes_per_pixel / gobs_in_x) * gobs_size * block_height;
+            const std::size_t x2 = x * bytes_per_pixel;
+            const std::size_t swizzle_offset = gob_address + table[x2 % gobs_in_x];
+            const std::size_t pixel_index = (x + y * width) * out_bytes_per_pixel;
+            data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
+            data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
+            std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
+        }
+    }
+}
+static void FastSwizzleData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
+                            u8* swizzled_data, u8* unswizzled_data, bool unswizzle,
+                            u32 block_height) {
    std::array<u8*, 2> data_ptrs;
    const std::size_t stride{width * bytes_per_pixel};
-    const std::size_t image_width_in_gobs{(stride + 63) / 64};
+    const std::size_t gobs_in_x = 64;
+    const std::size_t gobs_in_y = 8;
+    const std::size_t gobs_size = gobs_in_x * gobs_in_y;
+    const std::size_t image_width_in_gobs{(stride + gobs_in_x - 1) / gobs_in_x};
    const std::size_t copy_size{16};
    for (std::size_t y = 0; y < height; ++y) {
        const std::size_t initial_gob =
-            (y / (8 * block_height)) * 512 * block_height * image_width_in_gobs +
+            (y / (gobs_in_y * block_height)) * gobs_size * block_height * image_width_in_gobs +
-            (y % (8 * block_height) / 8) * 512;
+            (y % (gobs_in_y * block_height) / gobs_in_y) * gobs_size;
-        const std::size_t pixel_base{y * width * bytes_per_pixel};
+        const std::size_t pixel_base{y * width * out_bytes_per_pixel};
-        const auto& table = swizzle_table[y % 8];
+        const auto& table = fast_swizzle_table[y % gobs_in_y];
        for (std::size_t xb = 0; xb < stride; xb += copy_size) {
-            const std::size_t gob_address{initial_gob + (xb / 64) * 512 * block_height};
+            const std::size_t gob_address{initial_gob +
+                                          (xb / gobs_in_x) * gobs_size * block_height};
            const std::size_t swizzle_offset{gob_address + table[(xb / 16) % 4]};
-            const std::size_t pixel_index{xb + pixel_base};
+            const std::size_t out_x = xb * out_bytes_per_pixel / bytes_per_pixel;
+            const std::size_t pixel_index{out_x + pixel_base};
            data_ptrs[unswizzle] = swizzled_data + swizzle_offset;
            data_ptrs[!unswizzle] = unswizzled_data + pixel_index;
            std::memcpy(data_ptrs[0], data_ptrs[1], copy_size);
@@ -88,6 +97,17 @@ void FastSwizzleData(u32 width, u32 height, u32 bytes_per_pixel, u8* swizzled_da
    }
 }
+void CopySwizzledData(u32 width, u32 height, u32 bytes_per_pixel, u32 out_bytes_per_pixel,
+                      u8* swizzled_data, u8* unswizzled_data, bool unswizzle, u32 block_height) {
+    if (bytes_per_pixel % 3 != 0 && (width * bytes_per_pixel) % 16 == 0) {
+        FastSwizzleData(width, height, bytes_per_pixel, out_bytes_per_pixel, swizzled_data,
+                        unswizzled_data, unswizzle, block_height);
+    } else {
+        LegacySwizzleData(width, height, bytes_per_pixel, out_bytes_per_pixel, swizzled_data,
+                          unswizzled_data, unswizzle, block_height);
+    }
+}
 u32 BytesPerPixel(TextureFormat format) {
    switch (format) {
    case TextureFormat::DXT1:
@@ -134,13 +154,8 @@ u32 BytesPerPixel(TextureFormat format) {
 std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width,
                                 u32 height, u32 block_height) {
    std::vector<u8> unswizzled_data(width * height * bytes_per_pixel);
-    if (bytes_per_pixel % 3 != 0 && (width * bytes_per_pixel) % 16 == 0) {
+    CopySwizzledData(width / tile_size, height / tile_size, bytes_per_pixel, bytes_per_pixel,
-        FastSwizzleData(width / tile_size, height / tile_size, bytes_per_pixel,
+                     Memory::GetPointer(address), unswizzled_data.data(), true, block_height);
-                        Memory::GetPointer(address), unswizzled_data.data(), true, block_height);
-    } else {
-        CopySwizzledData(width / tile_size, height / tile_size, bytes_per_pixel, bytes_per_pixel,
-                         Memory::GetPointer(address), unswizzled_data.data(), true, block_height);
-    }
    return unswizzled_data;
 }
diff --git a/src/video_core/textures/texture.h b/src/video_core/textures/texture.h
index c2fb824b2..14aea4838 100644
--- a/src/video_core/textures/texture.h
+++ b/src/video_core/textures/texture.h
@@ -165,6 +165,8 @@ struct TICEntry {
        // High 16 bits of the pitch value
        BitField<0, 16, u32> pitch_high;
+        BitField<28, 4, u32> max_mip_level;
    };
    union {
        BitField<0, 16, u32> width_minus_1;
diff --git a/src/video_core/utils.h b/src/video_core/utils.h
index e0a14d48f..681919ae3 100644
--- a/src/video_core/utils.h
+++ b/src/video_core/utils.h
@@ -161,4 +161,26 @@ static inline void MortonCopyPixels128(u32 width, u32 height, u32 bytes_per_pixe
    }
 }
+static void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr,
+                          std::string extra_info = "") {
+    if (!GLAD_GL_KHR_debug) {
+        return; // We don't need to throw an error as this is just for debugging
+    }
+    const std::string nice_addr = fmt::format("0x{:016x}", addr);
+    std::string object_label;
+    switch (identifier) {
+    case GL_TEXTURE:
+        object_label = extra_info + "@" + nice_addr;
+        break;
+    case GL_PROGRAM:
+        object_label = "ShaderProgram@" + nice_addr;
+        break;
+    default:
+        object_label = fmt::format("Object(0x{:x})@{}", identifier, nice_addr);
+        break;
+    }
+    glObjectLabel(identifier, handle, -1, static_cast<const GLchar*>(object_label.c_str()));
+}
 } // namespace VideoCore
diff --git a/src/yuzu/debugger/wait_tree.cpp b/src/yuzu/debugger/wait_tree.cpp
index f2a7e23f0..a3b1fd357 100644
--- a/src/yuzu/debugger/wait_tree.cpp
+++ b/src/yuzu/debugger/wait_tree.cpp
@@ -15,6 +15,7 @@
 #include "core/hle/kernel/thread.h"
 #include "core/hle/kernel/timer.h"
 #include "core/hle/kernel/wait_object.h"
+#include "core/memory.h"
 WaitTreeItem::WaitTreeItem() = default;
 WaitTreeItem::~WaitTreeItem() = default;
diff --git a/src/yuzu/game_list.cpp b/src/yuzu/game_list.cpp
index 3b3b551bb..67890455a 100644
--- a/src/yuzu/game_list.cpp
+++ b/src/yuzu/game_list.cpp
@@ -26,10 +26,10 @@
 #include "yuzu/main.h"
 #include "yuzu/ui_settings.h"
-GameList::SearchField::KeyReleaseEater::KeyReleaseEater(GameList* gamelist) : gamelist{gamelist} {}
+GameListSearchField::KeyReleaseEater::KeyReleaseEater(GameList* gamelist) : gamelist{gamelist} {}
 // EventFilter in order to process systemkeys while editing the searchfield
-bool GameList::SearchField::KeyReleaseEater::eventFilter(QObject* obj, QEvent* event) {
+bool GameListSearchField::KeyReleaseEater::eventFilter(QObject* obj, QEvent* event) {
    // If it isn't a KeyRelease event then continue with standard event processing
    if (event->type() != QEvent::KeyRelease)
        return QObject::eventFilter(obj, event);
@@ -88,29 +88,21 @@ bool GameList::SearchField::KeyReleaseEater::eventFilter(QObject* obj, QEvent* e
    return QObject::eventFilter(obj, event);
 }
-void GameList::SearchField::setFilterResult(int visible, int total) {
+void GameListSearchField::setFilterResult(int visible, int total) {
-    QString result_of_text = tr("of");
+    label_filter_result->setText(tr("%1 of %n result(s)", "", total).arg(visible));
-    QString result_text;
-    if (total == 1) {
-        result_text = tr("result");
-    } else {
-        result_text = tr("results");
-    }
-    label_filter_result->setText(
-        QString("%1 %2 %3 %4").arg(visible).arg(result_of_text).arg(total).arg(result_text));
 }
-void GameList::SearchField::clear() {
+void GameListSearchField::clear() {
    edit_filter->setText("");
 }
-void GameList::SearchField::setFocus() {
+void GameListSearchField::setFocus() {
    if (edit_filter->isVisible()) {
        edit_filter->setFocus();
    }
 }
-GameList::SearchField::SearchField(GameList* parent) : QWidget{parent} {
+GameListSearchField::GameListSearchField(GameList* parent) : QWidget{parent} {
    KeyReleaseEater* keyReleaseEater = new KeyReleaseEater(parent);
    layout_filter = new QHBoxLayout;
    layout_filter->setMargin(8);
@@ -210,7 +202,7 @@ GameList::GameList(FileSys::VirtualFilesystem vfs, GMainWindow* parent)
    this->main_window = parent;
    layout = new QVBoxLayout;
    tree_view = new QTreeView;
-    search_field = new SearchField(this);
+    search_field = new GameListSearchField(this);
    item_model = new QStandardItemModel(tree_view);
    tree_view->setModel(item_model);
@@ -326,9 +318,14 @@ void GameList::PopupContextMenu(const QPoint& menu_location) {
    int row = item_model->itemFromIndex(item)->row();
    QStandardItem* child_file = item_model->invisibleRootItem()->child(row, COLUMN_NAME);
    u64 program_id = child_file->data(GameListItemPath::ProgramIdRole).toULongLong();
+    std::string path = child_file->data(GameListItemPath::FullPathRole).toString().toStdString();
    QMenu context_menu;
    QAction* open_save_location = context_menu.addAction(tr("Open Save Data Location"));
+    QAction* open_lfs_location = context_menu.addAction(tr("Open Mod Data Location"));
+    context_menu.addSeparator();
+    QAction* dump_romfs = context_menu.addAction(tr("Dump RomFS"));
+    QAction* copy_tid = context_menu.addAction(tr("Copy Title ID to Clipboard"));
    QAction* navigate_to_gamedb_entry = context_menu.addAction(tr("Navigate to GameDB entry"));
    open_save_location->setEnabled(program_id != 0);
@@ -337,6 +334,10 @@ void GameList::PopupContextMenu(const QPoint& menu_location) {
    connect(open_save_location, &QAction::triggered,
            [&]() { emit OpenFolderRequested(program_id, GameListOpenTarget::SaveData); });
+    connect(open_lfs_location, &QAction::triggered,
+            [&]() { emit OpenFolderRequested(program_id, GameListOpenTarget::ModData); });
+    connect(dump_romfs, &QAction::triggered, [&]() { emit DumpRomFSRequested(program_id, path); });
+    connect(copy_tid, &QAction::triggered, [&]() { emit CopyTIDRequested(program_id); });
    connect(navigate_to_gamedb_entry, &QAction::triggered,
            [&]() { emit NavigateToGamedbEntryRequested(program_id, compatibility_list); });
diff --git a/src/yuzu/game_list.h b/src/yuzu/game_list.h
index 2713e7b54..05e115e19 100644
--- a/src/yuzu/game_list.h
+++ b/src/yuzu/game_list.h
@@ -22,13 +22,17 @@
 #include "yuzu/compatibility_list.h"
 class GameListWorker;
+class GameListSearchField;
 class GMainWindow;
 namespace FileSys {
 class VfsFilesystem;
 }
-enum class GameListOpenTarget { SaveData };
+enum class GameListOpenTarget {
+    SaveData,
+    ModData,
+};
 class GameList : public QWidget {
    Q_OBJECT
@@ -43,33 +47,6 @@ public:
        COLUMN_COUNT, // Number of columns
    };
-    class SearchField : public QWidget {
-    public:
-        void setFilterResult(int visible, int total);
-        void clear();
-        void setFocus();
-        explicit SearchField(GameList* parent = nullptr);
-    private:
-        class KeyReleaseEater : public QObject {
-        public:
-            explicit KeyReleaseEater(GameList* gamelist);
-        private:
-            GameList* gamelist = nullptr;
-            QString edit_filter_text_old;
-        protected:
-            bool eventFilter(QObject* obj, QEvent* event) override;
-        };
-        QHBoxLayout* layout_filter = nullptr;
-        QTreeView* tree_view = nullptr;
-        QLabel* label_filter = nullptr;
-        QLineEdit* edit_filter = nullptr;
-        QLabel* label_filter_result = nullptr;
-        QToolButton* button_filter_close = nullptr;
-    };
    explicit GameList(std::shared_ptr<FileSys::VfsFilesystem> vfs, GMainWindow* parent = nullptr);
    ~GameList() override;
@@ -89,6 +66,8 @@ signals:
    void GameChosen(QString game_path);
    void ShouldCancelWorker();
    void OpenFolderRequested(u64 program_id, GameListOpenTarget target);
+    void DumpRomFSRequested(u64 program_id, const std::string& game_path);
+    void CopyTIDRequested(u64 program_id);
    void NavigateToGamedbEntryRequested(u64 program_id,
                                        const CompatibilityList& compatibility_list);
@@ -105,7 +84,7 @@ private:
    void RefreshGameDirectory();
    std::shared_ptr<FileSys::VfsFilesystem> vfs;
-    SearchField* search_field;
+    GameListSearchField* search_field;
    GMainWindow* main_window = nullptr;
    QVBoxLayout* layout = nullptr;
    QTreeView* tree_view = nullptr;
@@ -113,6 +92,8 @@ private:
    GameListWorker* current_worker = nullptr;
    QFileSystemWatcher* watcher = nullptr;
    CompatibilityList compatibility_list;
+    friend class GameListSearchField;
 };
 Q_DECLARE_METATYPE(GameListOpenTarget);
diff --git a/src/yuzu/game_list_p.h b/src/yuzu/game_list_p.h
index b6272d536..3db0e90da 100644
--- a/src/yuzu/game_list_p.h
+++ b/src/yuzu/game_list_p.h
@@ -16,6 +16,7 @@
 #include <QObject>
 #include <QStandardItem>
 #include <QString>
+#include <QWidget>
 #include "common/common_types.h"
 #include "common/logging/log.h"
@@ -68,7 +69,7 @@ public:
        if (!picture.loadFromData(picture_data.data(), static_cast<u32>(picture_data.size()))) {
            picture = GetDefaultIcon(size);
        }
-        picture = picture.scaled(size, size);
+        picture = picture.scaled(size, size, Qt::IgnoreAspectRatio, Qt::SmoothTransformation);
        setData(picture, Qt::DecorationRole);
    }
@@ -176,3 +177,42 @@ public:
        return data(SizeRole).toULongLong() < other.data(SizeRole).toULongLong();
    }
 };
+class GameList;
+class QHBoxLayout;
+class QTreeView;
+class QLabel;
+class QLineEdit;
+class QToolButton;
+class GameListSearchField : public QWidget {
+    Q_OBJECT
+public:
+    explicit GameListSearchField(GameList* parent = nullptr);
+    void setFilterResult(int visible, int total);
+    void clear();
+    void setFocus();
+private:
+    class KeyReleaseEater : public QObject {
+    public:
+        explicit KeyReleaseEater(GameList* gamelist);
+    private:
+        GameList* gamelist = nullptr;
+        QString edit_filter_text_old;
+    protected:
+        // EventFilter in order to process systemkeys while editing the searchfield
+        bool eventFilter(QObject* obj, QEvent* event) override;
+    };
+    QHBoxLayout* layout_filter = nullptr;
+    QTreeView* tree_view = nullptr;
+    QLabel* label_filter = nullptr;
+    QLineEdit* edit_filter = nullptr;
+    QLabel* label_filter_result = nullptr;
+    QToolButton* button_filter_close = nullptr;
+};
diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp
index 45bb1d1d1..27015d02c 100644
--- a/src/yuzu/main.cpp
+++ b/src/yuzu/main.cpp
@@ -7,6 +7,22 @@
 #include <memory>
 #include <thread>
+// VFS includes must be before glad as they will conflict with Windows file api, which uses defines.
+#include "core/file_sys/vfs.h"
+#include "core/file_sys/vfs_real.h"
+// These are wrappers to avoid the calls to CreateDirectory and CreateFile becuase of the Windows
+// defines.
+static FileSys::VirtualDir VfsFilesystemCreateDirectoryWrapper(
+    const FileSys::VirtualFilesystem& vfs, const std::string& path, FileSys::Mode mode) {
+    return vfs->CreateDirectory(path, mode);
+}
+static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::VirtualDir& dir,
+                                                          const std::string& path) {
+    return dir->CreateFile(path);
+}
 #include <fmt/ostream.h>
 #include <glad/glad.h>
@@ -30,16 +46,18 @@
 #include "common/telemetry.h"
 #include "core/core.h"
 #include "core/crypto/key_manager.h"
+#include "core/file_sys/bis_factory.h"
 #include "core/file_sys/card_image.h"
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/control_metadata.h"
 #include "core/file_sys/patch_manager.h"
 #include "core/file_sys/registered_cache.h"
+#include "core/file_sys/romfs.h"
 #include "core/file_sys/savedata_factory.h"
 #include "core/file_sys/submission_package.h"
-#include "core/file_sys/vfs_real.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/service/filesystem/filesystem.h"
+#include "core/hle/service/filesystem/fsp_ldr.h"
 #include "core/loader/loader.h"
 #include "core/perf_stats.h"
 #include "core/settings.h"
@@ -362,6 +380,8 @@ void GMainWindow::RestoreUIState() {
 void GMainWindow::ConnectWidgetEvents() {
    connect(game_list, &GameList::GameChosen, this, &GMainWindow::OnGameListLoadFile);
    connect(game_list, &GameList::OpenFolderRequested, this, &GMainWindow::OnGameListOpenFolder);
+    connect(game_list, &GameList::DumpRomFSRequested, this, &GMainWindow::OnGameListDumpRomFS);
+    connect(game_list, &GameList::CopyTIDRequested, this, &GMainWindow::OnGameListCopyTID);
    connect(game_list, &GameList::NavigateToGamedbEntryRequested, this,
            &GMainWindow::OnGameListNavigateToGamedbEntry);
@@ -602,9 +622,9 @@ void GMainWindow::BootGame(const QString& filename) {
    std::string title_name;
    const auto res = Core::System::GetInstance().GetGameName(title_name);
    if (res != Loader::ResultStatus::Success) {
-        const u64 program_id = Core::System::GetInstance().CurrentProcess()->program_id;
+        const u64 title_id = Core::System::GetInstance().CurrentProcess()->GetTitleID();
-        const auto [nacp, icon_file] = FileSys::PatchManager(program_id).GetControlMetadata();
+        const auto [nacp, icon_file] = FileSys::PatchManager(title_id).GetControlMetadata();
        if (nacp != nullptr)
            title_name = nacp->GetApplicationName();
@@ -713,6 +733,12 @@ void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target
                                                                program_id, user_id, 0);
        break;
    }
+    case GameListOpenTarget::ModData: {
+        open_target = "Mod Data";
+        const auto load_dir = FileUtil::GetUserPath(FileUtil::UserPath::LoadDir);
+        path = fmt::format("{}{:016X}", load_dir, program_id);
+        break;
+    }
    default:
        UNIMPLEMENTED();
    }
@@ -730,6 +756,120 @@ void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target
    QDesktopServices::openUrl(QUrl::fromLocalFile(qpath));
 }
+static std::size_t CalculateRomFSEntrySize(const FileSys::VirtualDir& dir, bool full) {
+    std::size_t out = 0;
+    for (const auto& subdir : dir->GetSubdirectories()) {
+        out += 1 + CalculateRomFSEntrySize(subdir, full);
+    }
+    return out + (full ? dir->GetFiles().size() : 0);
+}
+static bool RomFSRawCopy(QProgressDialog& dialog, const FileSys::VirtualDir& src,
+                         const FileSys::VirtualDir& dest, std::size_t block_size, bool full) {
+    if (src == nullptr || dest == nullptr || !src->IsReadable() || !dest->IsWritable())
+        return false;
+    if (dialog.wasCanceled())
+        return false;
+    if (full) {
+        for (const auto& file : src->GetFiles()) {
+            const auto out = VfsDirectoryCreateFileWrapper(dest, file->GetName());
+            if (!FileSys::VfsRawCopy(file, out, block_size))
+                return false;
+            dialog.setValue(dialog.value() + 1);
+            if (dialog.wasCanceled())
+                return false;
+        }
+    }
+    for (const auto& dir : src->GetSubdirectories()) {
+        const auto out = dest->CreateSubdirectory(dir->GetName());
+        if (!RomFSRawCopy(dialog, dir, out, block_size, full))
+            return false;
+        dialog.setValue(dialog.value() + 1);
+        if (dialog.wasCanceled())
+            return false;
+    }
+    return true;
+}
+void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_path) {
+    const auto path = fmt::format("{}{:016X}/romfs",
+                                  FileUtil::GetUserPath(FileUtil::UserPath::DumpDir), program_id);
+    const auto failed = [this, &path] {
+        QMessageBox::warning(this, tr("RomFS Extraction Failed!"),
+                             tr("There was an error copying the RomFS files or the user "
+                                "cancelled the operation."));
+        vfs->DeleteDirectory(path);
+    };
+    const auto loader = Loader::GetLoader(vfs->OpenFile(game_path, FileSys::Mode::Read));
+    if (loader == nullptr) {
+        failed();
+        return;
+    }
+    FileSys::VirtualFile file;
+    if (loader->ReadRomFS(file) != Loader::ResultStatus::Success) {
+        failed();
+        return;
+    }
+    const auto romfs =
+        loader->IsRomFSUpdatable()
+            ? FileSys::PatchManager(program_id).PatchRomFS(file, loader->ReadRomFSIVFCOffset())
+            : file;
+    const auto extracted = FileSys::ExtractRomFS(romfs, FileSys::RomFSExtractionType::Full);
+    if (extracted == nullptr) {
+        failed();
+        return;
+    }
+    const auto out = VfsFilesystemCreateDirectoryWrapper(vfs, path, FileSys::Mode::ReadWrite);
+    if (out == nullptr) {
+        failed();
+        return;
+    }
+    bool ok;
+    const auto res = QInputDialog::getItem(
+        this, tr("Select RomFS Dump Mode"),
+        tr("Please select the how you would like the RomFS dumped.<br>Full will copy all of the "
+           "files into the new directory while <br>skeleton will only create the directory "
+           "structure."),
+        {"Full", "Skeleton"}, 0, false, &ok);
+    if (!ok)
+        failed();
+    const auto full = res == "Full";
+    const auto entry_size = CalculateRomFSEntrySize(extracted, full);
+    QProgressDialog progress(tr("Extracting RomFS..."), tr("Cancel"), 0, entry_size, this);
+    progress.setWindowModality(Qt::WindowModal);
+    progress.setMinimumDuration(100);
+    if (RomFSRawCopy(progress, extracted, out, 0x400000, full)) {
+        progress.close();
+        QMessageBox::information(this, tr("RomFS Extraction Succeeded!"),
+                                 tr("The operation completed successfully."));
+        QDesktopServices::openUrl(QUrl::fromLocalFile(QString::fromStdString(path)));
+    } else {
+        progress.close();
+        failed();
+    }
+}
+void GMainWindow::OnGameListCopyTID(u64 program_id) {
+    QClipboard* clipboard = QGuiApplication::clipboard();
+    clipboard->setText(QString::fromStdString(fmt::format("{:016X}", program_id)));
+}
 void GMainWindow::OnGameListNavigateToGamedbEntry(u64 program_id,
                                                  const CompatibilityList& compatibility_list) {
    const auto it = FindMatchingCompatibilityEntry(compatibility_list, program_id);
@@ -790,7 +930,8 @@ void GMainWindow::OnMenuInstallToNAND() {
        return;
    }
-    const auto qt_raw_copy = [this](FileSys::VirtualFile src, FileSys::VirtualFile dest) {
+    const auto qt_raw_copy = [this](const FileSys::VirtualFile& src,
+                                    const FileSys::VirtualFile& dest, std::size_t block_size) {
        if (src == nullptr || dest == nullptr)
            return false;
        if (!dest->Resize(src->GetSize()))
@@ -914,11 +1055,21 @@ void GMainWindow::OnMenuInstallToNAND() {
            return;
        }
-        if (index >= 5)
+        // If index is equal to or past Game, add the jump in TitleType.
-            index += 0x7B;
+        if (index >= 5) {
+            index += static_cast<size_t>(FileSys::TitleType::Application) -
+                     static_cast<size_t>(FileSys::TitleType::FirmwarePackageB);
+        }
+        FileSys::InstallResult res;
+        if (index >= static_cast<size_t>(FileSys::TitleType::Application)) {
+            res = Service::FileSystem::GetUserNANDContents()->InstallEntry(
+                nca, static_cast<FileSys::TitleType>(index), false, qt_raw_copy);
+        } else {
+            res = Service::FileSystem::GetSystemNANDContents()->InstallEntry(
+                nca, static_cast<FileSys::TitleType>(index), false, qt_raw_copy);
+        }
-        const auto res = Service::FileSystem::GetUserNANDContents()->InstallEntry(
-            nca, static_cast<FileSys::TitleType>(index), false, qt_raw_copy);
        if (res == FileSys::InstallResult::Success) {
            success();
        } else if (res == FileSys::InstallResult::ErrorAlreadyExists) {
diff --git a/src/yuzu/main.h b/src/yuzu/main.h
index 552e3e61c..8ee9242b1 100644
--- a/src/yuzu/main.h
+++ b/src/yuzu/main.h
@@ -138,6 +138,8 @@ private slots:
    /// Called whenever a user selects a game in the game list widget.
    void OnGameListLoadFile(QString game_path);
    void OnGameListOpenFolder(u64 program_id, GameListOpenTarget target);
+    void OnGameListDumpRomFS(u64 program_id, const std::string& game_path);
+    void OnGameListCopyTID(u64 program_id);
    void OnGameListNavigateToGamedbEntry(u64 program_id,
                                         const CompatibilityList& compatibility_list);
    void OnMenuLoadFile();
diff --git a/src/yuzu_cmd/config.cpp b/src/yuzu_cmd/config.cpp
index 7ec1f5110..a478b0a56 100644
--- a/src/yuzu_cmd/config.cpp
+++ b/src/yuzu_cmd/config.cpp
@@ -120,7 +120,7 @@ void Config::ReadValues() {
                          sdl2_config->Get("Data Storage", "nand_directory",
                                           FileUtil::GetUserPath(FileUtil::UserPath::NANDDir)));
    FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir,
-                          sdl2_config->Get("Data Storage", "nand_directory",
+                          sdl2_config->Get("Data Storage", "sdmc_directory",
                                           FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir)));
    // System
diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp
index b1c364fbb..b2559b717 100644
--- a/src/yuzu_cmd/yuzu.cpp
+++ b/src/yuzu_cmd/yuzu.cpp
@@ -20,8 +20,10 @@
 #include "common/string_util.h"
 #include "common/telemetry.h"
 #include "core/core.h"
+#include "core/crypto/key_manager.h"
 #include "core/file_sys/vfs_real.h"
 #include "core/gdbstub/gdbstub.h"
+#include "core/hle/service/filesystem/filesystem.h"
 #include "core/loader/loader.h"
 #include "core/settings.h"
 #include "core/telemetry_session.h"
@@ -29,7 +31,6 @@
 #include "yuzu_cmd/emu_window/emu_window_sdl2.h"
 #include <getopt.h>
-#include "core/crypto/key_manager.h"
 #ifndef _MSC_VER
 #include <unistd.h>
 #endif
@@ -169,6 +170,7 @@ int main(int argc, char** argv) {
    Core::System& system{Core::System::GetInstance()};
    system.SetFilesystem(std::make_shared<FileSys::RealVfsFilesystem>());
+    Service::FileSystem::CreateFactories(system.GetFilesystem());
    SCOPE_EXIT({ system.Shutdown(); });