code: Add texture sampling option

* This replaces the nearest neighbour filter that shouldn't have existed in the first place
command_processor: Fix out-of-bounds float-uniform access (#7111 )
2023-11-23 01:22:47 +02:00 · 2023-11-03 03:35:52 -07:00 · 2023-11-03 03:35:32 -07:00 · 2023-11-02 17:33:40 -07:00 · 2023-11-02 17:19:00 -07:00 · 2023-11-01 17:58:02 -07:00
65 changed files with 1250 additions and 888 deletions
--- a/.ci/ios.sh
+++ b/.ci/ios.sh
@@ -1,5 +1,10 @@
 #!/bin/bash -ex

+# TODO: Work around pip install issues with Python 3.12 in the GitHub runner image.
+# See: https://github.com/actions/runner-images/issues/8709
+PYTHON_PATH=$(brew --prefix python@3.11)
+export PATH=$PYTHON_PATH/bin:$PYTHON_PATH/libexec/bin:$PATH
+
 mkdir build && cd build
 cmake .. -GNinja \
    -DCMAKE_BUILD_TYPE=Release \
--- a/.ci/linux.sh
+++ b/.ci/linux.sh
@@ -5,6 +5,9 @@ cmake .. -G Ninja \
    -DCMAKE_BUILD_TYPE=Release \
    -DCMAKE_C_COMPILER_LAUNCHER=ccache \
    -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
+    -DCMAKE_CXX_COMPILER=clang++ \
+    -DCMAKE_C_COMPILER=clang \
+    -DCMAKE_LINKER=/etc/bin/ld.lld \
    -DENABLE_QT_TRANSLATION=ON \
    -DCITRA_ENABLE_COMPATIBILITY_REPORTING=ON \
    -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON \
--- a/.ci/macos.sh
+++ b/.ci/macos.sh
@@ -1,5 +1,10 @@
 #!/bin/bash -ex

+# TODO: Work around pip install issues with Python 3.12 in the GitHub runner image.
+# See: https://github.com/actions/runner-images/issues/8709
+PYTHON_PATH=$(brew --prefix python@3.11)
+export PATH=$PYTHON_PATH/bin:$PYTHON_PATH/libexec/bin:$PATH
+
 mkdir build && cd build
 cmake .. -GNinja \
    -DCMAKE_BUILD_TYPE=Release \
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -59,7 +59,7 @@ jobs:
          name: ${{ env.OS }}-${{ env.TARGET }}
          path: artifacts/
  macos:
-    runs-on: macos-latest
+    runs-on: macos-13
    strategy:
      matrix:
        target: ["x86_64", "arm64"]
@@ -92,7 +92,7 @@ jobs:
          path: ${{ env.OS }}-${{ env.TARGET }}
          key: ${{ runner.os }}-${{ matrix.target }}-${{ github.sha }}-${{ github.run_id }}-${{ github.run_attempt }}
  macos-universal:
-    runs-on: macos-latest
+    runs-on: macos-13
    needs: macos
    env:
      OS: macos
@@ -158,7 +158,7 @@ jobs:
        if: ${{ matrix.target == 'msvc' }}
        with:
          vulkan-query-version: latest
-          vulkan-components: Glslang
+          vulkan-components: SPIRV-Tools, Glslang
          vulkan-use-cache: true
      - name: Set up MSYS2
        uses: msys2/setup-msys2@v2
@@ -234,7 +234,7 @@ jobs:
          name: ${{ env.OS }}-${{ env.TARGET }}
          path: src/android/app/artifacts/
  ios:
-    runs-on: macos-latest
+    runs-on: macos-13
    if: ${{ !startsWith(github.ref, 'refs/tags/') }}
    env:
      CCACHE_DIR: ${{ github.workspace }}/.ccache
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -17,6 +17,12 @@ include(CMakeDependentOption)

 project(citra LANGUAGES C CXX ASM)

+# Some submodules like to pick their own default build type if not specified.
+# Make sure we default to Release build type always, unless the generator has custom types.
+if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
+    set(CMAKE_BUILD_TYPE "Release" CACHE STRING "Choose the type of build." FORCE)
+endif()
+
 if (APPLE)
    # Silence warnings on empty objects, for example when platform-specific code is #ifdef'd out.
    set(CMAKE_C_ARCHIVE_CREATE "<CMAKE_AR> Scr <TARGET> <LINK_FLAGS> <OBJECTS>")
@@ -86,24 +92,7 @@ option(ENABLE_LTO "Enable link time optimization" ${DEFAULT_ENABLE_LTO})
 option(CITRA_USE_PRECOMPILED_HEADERS "Use precompiled headers" ON)
 option(CITRA_WARNINGS_AS_ERRORS "Enable warnings as errors" ON)

-# System library options
-CMAKE_DEPENDENT_OPTION(USE_SYSTEM_QT "Use the system Qt lib (instead of the bundled one)" OFF "ENABLE_QT;MSVC OR APPLE" ON)
-CMAKE_DEPENDENT_OPTION(USE_SYSTEM_MOLTENVK "Use the system MoltenVK lib (instead of the bundled one)" OFF "APPLE" OFF)
-option(USE_SYSTEM_SDL2 "Use the system SDL2 lib (instead of the bundled one)" OFF)
-option(USE_SYSTEM_BOOST "Use the system Boost libs (instead of the bundled ones)" OFF)
-option(USE_SYSTEM_OPENSSL "Use the system OpenSSL libs (instead of the bundled LibreSSL)" OFF)
-option(USE_SYSTEM_LIBUSB "Use the system libusb (instead of the bundled libusb)" OFF)
-option(USE_SYSTEM_CPP_JWT "Use the system cpp-jwt (instead of the bundled one)" OFF)
-option(USE_SYSTEM_SOUNDTOUCH "Use the system SoundTouch (instead of the bundled one)" OFF)
-option(USE_SYSTEM_CPP_HTTPLIB "Use the system cpp-httplib (instead of the bundled one)" OFF)
-option(USE_SYSTEM_JSON "Use the system JSON (nlohmann-json3) package (instead of the bundled one)" OFF)
-option(USE_SYSTEM_DYNARMIC "Use the system dynarmic (instead of the bundled one)" OFF)
-option(USE_SYSTEM_FMT "Use the system fmt (instead of the bundled one)" OFF)
-option(USE_SYSTEM_XBYAK "Use the system xbyak (instead of the bundled one)" OFF)
-option(USE_SYSTEM_INIH "Use the system inih (instead of the bundled one)" OFF)
-option(USE_SYSTEM_FDK_AAC_HEADERS "Use the system fdk-aac headers (instead of the bundled one)" OFF)
-option(USE_SYSTEM_FFMPEG_HEADERS "Use the system FFmpeg headers (instead of the bundled one)" OFF)
-option(USE_SYSTEM_GLSLANG "Use the system glslang and SPIR-V libraries (instead of the bundled ones)" OFF)
+include(CitraHandleSystemLibs)

 if (CITRA_USE_PRECOMPILED_HEADERS)
    message(STATUS "Using Precompiled Headers.")
@@ -247,7 +236,7 @@ find_package(Threads REQUIRED)

 if (ENABLE_QT)
    if (NOT USE_SYSTEM_QT)
-        download_qt(6.5.1)
+        download_qt(6.6.0)
    endif()

    find_package(Qt6 REQUIRED COMPONENTS Widgets Multimedia Concurrent)
--- a/dist/qt_themes/colorful_dark/icons/index.theme
+++ b/dist/qt_themes/colorful_dark/icons/index.theme
@@ -2,7 +2,13 @@
 Name=colorful_dark
 Comment=Colorful theme (Dark style)
 Inherits=default
-Directories=16x16
- 
+Directories=16x16,48x48,256x256
+
 [16x16]
 Size=16
+
+[48x48]
+Size=48
+
+[256x256]
+Size=256
--- a/dist/qt_themes/colorful_midnight_blue/icons/index.theme
+++ b/dist/qt_themes/colorful_midnight_blue/icons/index.theme
@@ -2,7 +2,13 @@
 Name=colorful_midnight_blue
 Comment=Colorful theme (Midnight Blue style)
 Inherits=default
-Directories=16x16
+Directories=16x16,48x48,256x256

 [16x16]
 Size=16
+
+[48x48]
+Size=48
+
+[256x256]
+Size=256
--- a/externals/CMakeLists.txt
+++ b/externals/CMakeLists.txt
@@ -34,6 +34,10 @@ if (NOT USE_SYSTEM_BOOST)
    )
    target_link_libraries(boost_iostreams PUBLIC boost)
 # Add additional boost libs here; remember to ALIAS them in the root CMakeLists!
+else()
+    unset(BOOST_ROOT CACHE)
+    unset(Boost_INCLUDE_DIR CACHE)
+    set(Boost_NO_SYSTEM_PATHS OFF CACHE BOOL "" FORCE)
 endif()

 # Catch2
@@ -42,11 +46,17 @@ set(CATCH_INSTALL_EXTRAS OFF CACHE BOOL "")
 add_subdirectory(catch2)

 # Crypto++
-set(CRYPTOPP_BUILD_DOCUMENTATION OFF CACHE BOOL "")
-set(CRYPTOPP_BUILD_TESTING OFF CACHE BOOL "")
-set(CRYPTOPP_INSTALL OFF CACHE BOOL "")
-set(CRYPTOPP_SOURCES "${CMAKE_SOURCE_DIR}/externals/cryptopp" CACHE STRING "")
-add_subdirectory(cryptopp-cmake)
+if(USE_SYSTEM_CRYPTOPP)
+    find_package(cryptopp REQUIRED)
+    add_library(cryptopp INTERFACE)
+    target_link_libraries(cryptopp INTERFACE cryptopp::cryptopp)
+else()
+    set(CRYPTOPP_BUILD_DOCUMENTATION OFF CACHE BOOL "")
+    set(CRYPTOPP_BUILD_TESTING OFF CACHE BOOL "")
+    set(CRYPTOPP_INSTALL OFF CACHE BOOL "")
+    set(CRYPTOPP_SOURCES "${CMAKE_SOURCE_DIR}/externals/cryptopp" CACHE STRING "")
+    add_subdirectory(cryptopp-cmake)
+endif()

 # dds-ktx
 add_library(dds-ktx INTERFACE)
@@ -207,22 +217,47 @@ if (ENABLE_LIBUSB AND NOT USE_SYSTEM_LIBUSB)
 endif()

 # Zstandard
-set(ZSTD_LEGACY_SUPPORT OFF)
-set(ZSTD_BUILD_PROGRAMS OFF)
-set(ZSTD_BUILD_SHARED OFF)
-add_subdirectory(zstd/build/cmake EXCLUDE_FROM_ALL)
-target_include_directories(libzstd_static INTERFACE $<BUILD_INTERFACE:${CMAKE_SOURCE_DIR}/externals/zstd/lib>)
+if(USE_SYSTEM_ZSTD)
+    find_package(zstd REQUIRED)
+    add_library(zstd INTERFACE)
+    if(TARGET zstd::libzstd_shared)
+        message(STATUS "Found system Zstandard")
+    endif()
+    target_link_libraries(zstd INTERFACE zstd::libzstd_shared)
+else()
+    set(ZSTD_LEGACY_SUPPORT OFF)
+    set(ZSTD_BUILD_PROGRAMS OFF)
+    set(ZSTD_BUILD_SHARED OFF)
+    add_subdirectory(zstd/build/cmake EXCLUDE_FROM_ALL)
+    target_include_directories(libzstd_static INTERFACE $<BUILD_INTERFACE:${CMAKE_SOURCE_DIR}/externals/zstd/lib>)
+    add_library(zstd ALIAS libzstd_static)
+endif()

 # ENet
-add_subdirectory(enet)
-target_include_directories(enet INTERFACE ./enet/include)
+if(USE_SYSTEM_ENET)
+    find_package(libenet REQUIRED)
+    add_library(enet INTERFACE)
+    target_link_libraries(enet INTERFACE libenet::libenet)
+else()
+    add_subdirectory(enet)
+    target_include_directories(enet INTERFACE ./enet/include)
+endif()

 # Cubeb
 if (ENABLE_CUBEB)
-    set(BUILD_TESTS OFF CACHE BOOL "")
-    set(BUILD_TOOLS OFF CACHE BOOL "")
-    set(BUNDLE_SPEEX ON CACHE BOOL "")
-    add_subdirectory(cubeb EXCLUDE_FROM_ALL)
+    if(USE_SYSTEM_CUBEB)
+        find_package(cubeb REQUIRED)
+        add_library(cubeb INTERFACE)
+        target_link_libraries(cubeb INTERFACE cubeb::cubeb)
+        if(TARGET cubeb::cubeb)
+            message(STATUS "Found system cubeb")
+        endif()
+    else()
+        set(BUILD_TESTS OFF CACHE BOOL "")
+        set(BUILD_TOOLS OFF CACHE BOOL "")
+        set(BUNDLE_SPEEX ON CACHE BOOL "")
+        add_subdirectory(cubeb EXCLUDE_FROM_ALL)
+    endif()
 endif()

 # DiscordRPC
--- a/externals/cmake-modules/CitraHandleSystemLibs.cmake
+++ b/externals/cmake-modules/CitraHandleSystemLibs.cmake
@@ -0,0 +1,88 @@
+option(USE_SYSTEM_LIBS "Use system libraries over bundled ones" OFF)
+
+# System library options
+CMAKE_DEPENDENT_OPTION(USE_SYSTEM_QT "Use the system Qt lib (instead of the bundled one)" OFF "ENABLE_QT;MSVC OR APPLE" ON)
+CMAKE_DEPENDENT_OPTION(USE_SYSTEM_MOLTENVK "Use the system MoltenVK lib (instead of the bundled one)" OFF "APPLE" OFF)
+option(USE_SYSTEM_SDL2 "Use the system SDL2 lib (instead of the bundled one)" OFF)
+option(USE_SYSTEM_BOOST "Use the system Boost libs (instead of the bundled ones)" OFF)
+option(USE_SYSTEM_OPENSSL "Use the system OpenSSL libs (instead of the bundled LibreSSL)" OFF)
+option(USE_SYSTEM_LIBUSB "Use the system libusb (instead of the bundled libusb)" OFF)
+option(USE_SYSTEM_CPP_JWT "Use the system cpp-jwt (instead of the bundled one)" OFF)
+option(USE_SYSTEM_SOUNDTOUCH "Use the system SoundTouch (instead of the bundled one)" OFF)
+option(USE_SYSTEM_CPP_HTTPLIB "Use the system cpp-httplib (instead of the bundled one)" OFF)
+option(USE_SYSTEM_JSON "Use the system JSON (nlohmann-json3) package (instead of the bundled one)" OFF)
+option(USE_SYSTEM_DYNARMIC "Use the system dynarmic (instead of the bundled one)" OFF)
+option(USE_SYSTEM_FMT "Use the system fmt (instead of the bundled one)" OFF)
+option(USE_SYSTEM_XBYAK "Use the system xbyak (instead of the bundled one)" OFF)
+option(USE_SYSTEM_INIH "Use the system inih (instead of the bundled one)" OFF)
+option(USE_SYSTEM_FDK_AAC_HEADERS "Use the system fdk-aac headers (instead of the bundled one)" OFF)
+option(USE_SYSTEM_FFMPEG_HEADERS "Use the system FFmpeg headers (instead of the bundled one)" OFF)
+option(USE_SYSTEM_GLSLANG "Use the system glslang and SPIR-V libraries (instead of the bundled ones)" OFF)
+option(USE_SYSTEM_ZSTD "Use the system Zstandard library (instead of the bundled one)" OFF)
+option(USE_SYSTEM_ENET "Use the system libenet (instead of the bundled one)" OFF)
+option(USE_SYSTEM_CRYPTOPP "Use the system cryptopp (instead of the bundled one)" OFF)
+option(USE_SYSTEM_CUBEB "Use the system cubeb (instead of the bundled one)" OFF)
+
+# Qt and MoltenVK are handled separately
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_SDL2 "Disable system SDL2" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_BOOST "Disable system Boost" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_OPENSSL "Disable system OpenSSL" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_LIBUSB "Disable system LibUSB" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_CPP_JWT "Disable system cpp-jwt" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_SOUNDTOUCH "Disable system SoundTouch" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_CPP_HTTPLIB "Disable system cpp-httplib" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_JSON "Disable system JSON" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_DYNARMIC "Disable system Dynarmic" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_FMT "Disable system fmt" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_XBYAK "Disable system xbyak" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_INIH "Disable system inih" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_FDK_AAC_HEADERS "Disable system fdk_aac" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_FFMPEG_HEADERS "Disable system ffmpeg" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_GLSLANG "Disable system glslang" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_ZSTD "Disable system Zstandard" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_ENET "Disable system libenet" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_CRYPTOPP "Disable system cryptopp" OFF "USE_SYSTEM_LIBS" OFF)
+CMAKE_DEPENDENT_OPTION(DISABLE_SYSTEM_CUBEB "Disable system cubeb" OFF "USE_SYSTEM_LIBS" OFF)
+
+set(LIB_VAR_LIST
+    SDL2
+    BOOST
+    OPENSSL
+    LIBUSB
+    CPP_JWT
+    SOUNDTOUCH
+    CPP_HTTPLIB
+    JSON
+    DYNARMIC
+    FMT
+    XBYAK
+    INIH
+    FDK_AAC_HEADERS
+    FFMPEG_HEADERS
+    GLSLANG
+    ZSTD
+    ENET
+    CRYPTOPP
+    CUBEB
+    )
+
+# First, check that USE_SYSTEM_XXX is not used with USE_SYSTEM_LIBS
+
+if(USE_SYSTEM_LIBS)
+    foreach(CURRENT_LIB IN LISTS LIB_VAR_LIST)
+        if(USE_SYSTEM_${CURRENT_LIB})
+            unset(USE_SYSTEM_${CURRENT_LIB})
+        endif()
+    endforeach()
+
+    # Next, set which libraries to use
+
+    foreach(CURRENT_LIB IN LISTS LIB_VAR_LIST)
+        if(NOT DISABLE_SYSTEM_${CURRENT_LIB})
+            set(USE_SYSTEM_${CURRENT_LIB} ON CACHE BOOL "Using system ${CURRENT_LIB}" FORCE)
+        else()
+            # Explicitly disable this in case of multiple CMake invocations
+            set(USE_SYSTEM_${CURRENT_LIB} OFF CACHE BOOL "Using system ${CURRENT_LIB}" FORCE)
+        endif()
+    endforeach()
+endif()
--- a/externals/cmake-modules/Findcryptopp.cmake
+++ b/externals/cmake-modules/Findcryptopp.cmake
@@ -0,0 +1,36 @@
+if(NOT CRYPTOPP_FOUND)
+    pkg_check_modules(CRYPTOPP_TMP libcrypto++)
+
+    find_path(CRYPTOPP_INCLUDE_DIRS NAMES cryptlib.h
+            PATHS
+            ${CRYPTOPP_TMP_INCLUDE_DIRS}
+            /usr/include
+            /usr/include/crypto++
+            /usr/local/include
+            /usr/local/include/crypto++
+            )
+
+    find_library(CRYPTOPP_LIBRARY_DIRS NAMES crypto++
+            PATHS
+            ${CRYPTOPP_TMP_LIBRARY_DIRS}
+            /usr/lib
+            /usr/locallib
+            )
+
+    if(CRYPTOPP_INCLUDE_DIRS AND CRYPTOPP_LIBRARY_DIRS)
+        set(CRYPTOPP_FOUND TRUE CACHE INTERNAL "Found cryptopp")
+        message(STATUS "Found cryptopp: ${CRYPTOPP_LIBRARY_DIRS}, ${CRYPTOPP_INCLUDE_DIRS}")
+    else()
+        set(CRYPTOPP_FOUND FALSE CACHE INTERNAL "Found cryptopp")
+        message(STATUS "Cryptopp not found.")
+    endif()
+endif()
+
+if(CRYPTOPP_FOUND AND NOT TARGET cryptopp::cryptopp)
+    add_library(cryptopp::cryptopp UNKNOWN IMPORTED)
+    set_target_properties(cryptopp::cryptopp PROPERTIES
+        INCLUDE_DIRECTORIES ${CRYPTOPP_INCLUDE_DIRS}
+        INTERFACE_LINK_LIBRARIES ${CRYPTOPP_LIBRARY_DIRS}
+        IMPORTED_LOCATION ${CRYPTOPP_LIBRARY_DIRS}
+        )
+endif()
--- a/externals/cmake-modules/Findlibenet.cmake
+++ b/externals/cmake-modules/Findlibenet.cmake
@@ -0,0 +1,34 @@
+if(NOT libenet_FOUND)
+    pkg_check_modules(ENET_TMP libenet)
+
+    find_path(libenet_INCLUDE_DIRS NAMES enet.h PATH_SUFFIXES enet
+            PATHS
+            ${ENET_TMP_INCLUDE_DIRS}
+            /usr/include
+            /usr/local/include
+            )
+
+    find_library(libenet_LIBRARY_DIRS NAMES enet
+            PATHS
+            ${ENET_TMP_LIBRARY_DIRS}
+            /usr/lib
+            /usr/local/lib
+            )
+
+    if(libenet_INCLUDE_DIRS AND libenet_LIBRARY_DIRS)
+        set(libenet_FOUND TRUE CACHE INTERNAL "Found libenet")
+        message(STATUS "Found libenet ${libenet_LIBRARY_DIRS}, ${libenet_INCLUDE_DIRS}")
+    else()
+        set(libenet_FOUND FALSE CACHE INTERNAL "Found libenet")
+        message(STATUS "Libenet not found.")
+    endif()
+endif()
+
+if(libenet_FOUND AND NOT TARGET libenet::libenet)
+    add_library(libenet::libenet UNKNOWN IMPORTED)
+    set_target_properties(libenet::libenet PROPERTIES
+            INCLUDE_DIRECTORIES ${libenet_INCLUDE_DIRS}
+            INTERFACE_LINK_LIBRARIES ${libenet_LIBRARY_DIRS}
+            IMPORTED_LOCATION ${libenet_LIBRARY_DIRS}
+            )
+endif()
--- a/src/audio_core/dsp_interface.cpp
+++ b/src/audio_core/dsp_interface.cpp
@@ -66,18 +66,22 @@ void DspInterface::OutputSample(std::array<s16, 2> sample) {
 }

 void DspInterface::OutputCallback(s16* buffer, std::size_t num_frames) {
-    std::size_t frames_written;
+    std::size_t frames_written = 0;
    if (perform_time_stretching) {
        const std::vector<s16> in{fifo.Pop()};
        const std::size_t num_in{in.size() / 2};
        frames_written = time_stretcher.Process(in.data(), num_in, buffer, num_frames);
-    } else if (flushing_time_stretcher) {
-        time_stretcher.Flush();
-        frames_written = time_stretcher.Process(nullptr, 0, buffer, num_frames);
-        frames_written += fifo.Pop(buffer, num_frames - frames_written);
-        flushing_time_stretcher = false;
    } else {
-        frames_written = fifo.Pop(buffer, num_frames);
+        if (flushing_time_stretcher) {
+            time_stretcher.Flush();
+            frames_written = time_stretcher.Process(nullptr, 0, buffer, num_frames);
+            flushing_time_stretcher = false;
+
+            // Make sure any frames that did not fit are cleared from the time stretcher,
+            // so that they do not bleed into the next time the stretcher is enabled.
+            time_stretcher.Clear();
+        }
+        frames_written += fifo.Pop(buffer, num_frames - frames_written);
    }

    if (frames_written > 0) {
--- a/src/audio_core/time_stretch.cpp
+++ b/src/audio_core/time_stretch.cpp
@@ -18,8 +18,7 @@

 namespace AudioCore {

-TimeStretcher::TimeStretcher()
-    : sample_rate(native_sample_rate), sound_touch(std::make_unique<soundtouch::SoundTouch>()) {
+TimeStretcher::TimeStretcher() : sound_touch(std::make_unique<soundtouch::SoundTouch>()) {
    sound_touch->setChannels(2);
    sound_touch->setSampleRate(native_sample_rate);
    sound_touch->setPitch(1.0);
@@ -30,16 +29,15 @@ TimeStretcher::~TimeStretcher() = default;

 void TimeStretcher::SetOutputSampleRate(unsigned int sample_rate) {
    sound_touch->setSampleRate(sample_rate);
-    sample_rate = native_sample_rate;
 }

 std::size_t TimeStretcher::Process(const s16* in, std::size_t num_in, s16* out,
                                   std::size_t num_out) {
-    const double time_delta = static_cast<double>(num_out) / sample_rate; // seconds
+    const double time_delta = static_cast<double>(num_out) / native_sample_rate; // seconds
    double current_ratio = static_cast<double>(num_in) / static_cast<double>(num_out);

    const double max_latency = 0.25; // seconds
-    const double max_backlog = sample_rate * max_latency;
+    const double max_backlog = native_sample_rate * max_latency;
    const double backlog_fullness = sound_touch->numSamples() / max_backlog;
    if (backlog_fullness > 4.0) {
        // Too many samples in backlog: Don't push anymore on
--- a/src/audio_core/time_stretch.h
+++ b/src/audio_core/time_stretch.h
@@ -34,7 +34,6 @@ public:
    void Flush();

 private:
-    unsigned int sample_rate;
    std::unique_ptr<soundtouch::SoundTouch> sound_touch;
    double stretch_ratio = 1.0;
 };
--- a/src/citra_qt/bootmanager.cpp
+++ b/src/citra_qt/bootmanager.cpp
@@ -669,9 +669,9 @@ void GRenderWindow::CaptureScreenshot(u32 res_scale, const QString& screenshot_p
    screenshot_image = QImage(QSize(layout.width, layout.height), QImage::Format_RGB32);
    system.Renderer().RequestScreenshot(
        screenshot_image.bits(),
-        [this, screenshot_path] {
+        [this, screenshot_path](bool invert_y) {
            const std::string std_screenshot_path = screenshot_path.toStdString();
-            if (screenshot_image.mirrored(false, true).save(screenshot_path)) {
+            if (screenshot_image.mirrored(false, invert_y).save(screenshot_path)) {
                LOG_INFO(Frontend, "Screenshot saved to \"{}\"", std_screenshot_path);
            } else {
                LOG_ERROR(Frontend, "Failed to save screenshot to \"{}\"", std_screenshot_path);
--- a/src/citra_qt/configuration/configure_enhancements.ui
+++ b/src/citra_qt/configuration/configure_enhancements.ui
@@ -184,11 +184,6 @@
             <string>Bicubic</string>
            </property>
           </item>
-           <item>
-            <property name="text">
-             <string>Nearest Neighbor</string>
-            </property>
-           </item>
           <item>
            <property name="text">
             <string>ScaleForce</string>
@@ -199,11 +194,11 @@
             <string>xBRZ</string>
            </property>
           </item>
-            <item>
-              <property name="text">
-                <string>MMPX</string>
-              </property>
-            </item>
+           <item>
+            <property name="text">
+             <string>MMPX</string>
+            </property>
+           </item>
          </widget>
         </item>
        </layout>
--- a/src/citra_qt/configuration/configure_graphics.cpp
+++ b/src/citra_qt/configuration/configure_graphics.cpp
@@ -71,11 +71,17 @@ void ConfigureGraphics::SetConfiguration() {
                                          !Settings::values.physical_device.UsingGlobal());
        ConfigurationShared::SetPerGameSetting(ui->physical_device_combo,
                                               &Settings::values.physical_device);
+        ConfigurationShared::SetPerGameSetting(ui->texture_sampling_combobox,
+                                               &Settings::values.texture_sampling);
+        ConfigurationShared::SetHighlight(ui->widget_texture_sampling,
+                                          !Settings::values.texture_sampling.UsingGlobal());
    } else {
        ui->graphics_api_combo->setCurrentIndex(
            static_cast<int>(Settings::values.graphics_api.GetValue()));
        ui->physical_device_combo->setCurrentIndex(
            static_cast<int>(Settings::values.physical_device.GetValue()));
+        ui->texture_sampling_combobox->setCurrentIndex(
+            static_cast<int>(Settings::values.texture_sampling.GetValue()));
    }

    ui->toggle_hw_shader->setChecked(Settings::values.use_hw_shader.GetValue());
@@ -106,6 +112,8 @@ void ConfigureGraphics::ApplyConfiguration() {
                                             use_hw_shader);
    ConfigurationShared::ApplyPerGameSetting(&Settings::values.shaders_accurate_mul,
                                             ui->toggle_accurate_mul, shaders_accurate_mul);
+    ConfigurationShared::ApplyPerGameSetting(&Settings::values.texture_sampling,
+                                             ui->texture_sampling_combobox);
    ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_disk_shader_cache,
                                             ui->toggle_disk_shader_cache, use_disk_shader_cache);
    ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_vsync_new, ui->toggle_vsync_new,
@@ -132,6 +140,7 @@ void ConfigureGraphics::SetupPerGameUI() {
                                         Settings::values.use_vsync_new.UsingGlobal());
        ui->toggle_async_shaders->setEnabled(
            Settings::values.async_shader_compilation.UsingGlobal());
+        ui->widget_texture_sampling->setEnabled(Settings::values.texture_sampling.UsingGlobal());
        ui->toggle_async_present->setEnabled(Settings::values.async_presentation.UsingGlobal());
        ui->graphics_api_combo->setEnabled(Settings::values.graphics_api.UsingGlobal());
        ui->physical_device_combo->setEnabled(Settings::values.physical_device.UsingGlobal());
@@ -148,6 +157,10 @@ void ConfigureGraphics::SetupPerGameUI() {
        ui->physical_device_combo, ui->physical_device_group,
        static_cast<u32>(Settings::values.physical_device.GetValue(true)));

+    ConfigurationShared::SetColoredComboBox(
+        ui->texture_sampling_combobox, ui->widget_texture_sampling,
+        static_cast<int>(Settings::values.texture_sampling.GetValue(true)));
+
    ConfigurationShared::SetColoredTristate(ui->toggle_hw_shader, Settings::values.use_hw_shader,
                                            use_hw_shader);
    ConfigurationShared::SetColoredTristate(
--- a/src/citra_qt/configuration/configure_graphics.ui
+++ b/src/citra_qt/configuration/configure_graphics.ui
@@ -212,6 +212,53 @@
      <string>Advanced</string>
     </property>
     <layout class="QVBoxLayout" name="verticalLayout_2">
+      <item>
+       <widget class="QWidget" name="widget_texture_sampling" native="true">
+        <layout class="QHBoxLayout" name="horizontalLayout_2">
+         <property name="leftMargin">
+          <number>0</number>
+         </property>
+         <property name="topMargin">
+          <number>0</number>
+         </property>
+         <property name="rightMargin">
+          <number>0</number>
+         </property>
+         <property name="bottomMargin">
+          <number>0</number>
+         </property>
+         <item>
+          <widget class="QLabel" name="texture_sampling_label">
+           <property name="toolTip">
+            <string>&lt;html&gt;&lt;head/&gt;&lt;body&gt;&lt;p&gt;Overrides the sampling filter used by games. This can be useful in certain cases with poorly behaved games when upscaling. If unsure set this to Game Controlled&lt;/p&gt;&lt;/body&gt;&lt;/html&gt;</string>
+           </property>
+           <property name="text">
+            <string>Texture Sampling</string>
+           </property>
+          </widget>
+         </item>
+         <item>
+          <widget class="QComboBox" name="texture_sampling_combobox">
+           <item>
+            <property name="text">
+             <string>Game Controlled</string>
+            </property>
+           </item>
+           <item>
+            <property name="text">
+             <string>Nearest Neighbor</string>
+            </property>
+           </item>
+           <item>
+            <property name="text">
+             <string>Linear</string>
+            </property>
+           </item>
+          </widget>
+         </item>
+        </layout>
+       </widget>
+      </item>
      <item>
       <widget class="QCheckBox" name="toggle_disk_shader_cache">
        <property name="toolTip">
--- a/src/citra_qt/game_list.cpp
+++ b/src/citra_qt/game_list.cpp
@@ -258,13 +258,15 @@ void GameList::OnUpdateThemedIcons() {
    for (int i = 0; i < item_model->invisibleRootItem()->rowCount(); i++) {
        QStandardItem* child = item_model->invisibleRootItem()->child(i);

+        const int icon_size = IconSizes.at(UISettings::values.game_list_icon_size.GetValue());
        switch (child->data(GameListItem::TypeRole).value<GameListItemType>()) {
        case GameListItemType::InstalledDir:
-            child->setData(QIcon::fromTheme(QStringLiteral("sd_card")).pixmap(48),
+            child->setData(QIcon::fromTheme(QStringLiteral("sd_card")).pixmap(icon_size),
                           Qt::DecorationRole);
            break;
        case GameListItemType::SystemDir:
-            child->setData(QIcon::fromTheme(QStringLiteral("chip")).pixmap(48), Qt::DecorationRole);
+            child->setData(QIcon::fromTheme(QStringLiteral("chip")).pixmap(icon_size),
+                           Qt::DecorationRole);
            break;
        case GameListItemType::CustomDir: {
            const UISettings::GameDir& game_dir =
@@ -272,11 +274,12 @@ void GameList::OnUpdateThemedIcons() {
            const QString icon_name = QFileInfo::exists(game_dir.path)
                                          ? QStringLiteral("folder")
                                          : QStringLiteral("bad_folder");
-            child->setData(QIcon::fromTheme(icon_name).pixmap(48), Qt::DecorationRole);
+            child->setData(QIcon::fromTheme(icon_name).pixmap(icon_size), Qt::DecorationRole);
            break;
        }
        case GameListItemType::AddDir:
-            child->setData(QIcon::fromTheme(QStringLiteral("plus")).pixmap(48), Qt::DecorationRole);
+            child->setData(QIcon::fromTheme(QStringLiteral("plus")).pixmap(icon_size),
+                           Qt::DecorationRole);
            break;
        default:
            break;
--- a/src/citra_qt/main.cpp
+++ b/src/citra_qt/main.cpp
@@ -2741,7 +2741,10 @@ void GMainWindow::filterBarSetChecked(bool state) {
 }

 void GMainWindow::UpdateUITheme() {
-    const QString default_icons = QStringLiteral(":/icons/default");
+    const QString icons_base_path = QStringLiteral(":/icons/");
+    const QString default_theme = QStringLiteral("default");
+    const QString default_theme_path = icons_base_path + default_theme;
+
    const QString& current_theme = UISettings::values.theme;
    const bool is_default_theme = current_theme == QString::fromUtf8(UISettings::themes[0].second);
    QStringList theme_paths(default_theme_paths);
@@ -2759,8 +2762,8 @@ void GMainWindow::UpdateUITheme() {
            qApp->setStyleSheet({});
            setStyleSheet({});
        }
-        theme_paths.append(default_icons);
-        QIcon::setThemeName(default_icons);
+        theme_paths.append(default_theme_path);
+        QIcon::setThemeName(default_theme);
    } else {
        const QString theme_uri(QLatin1Char{':'} + current_theme + QStringLiteral("/style.qss"));
        QFile f(theme_uri);
@@ -2772,9 +2775,9 @@ void GMainWindow::UpdateUITheme() {
            LOG_ERROR(Frontend, "Unable to set style, stylesheet file not found");
        }

-        const QString theme_name = QStringLiteral(":/icons/") + current_theme;
-        theme_paths.append({default_icons, theme_name});
-        QIcon::setThemeName(theme_name);
+        const QString current_theme_path = icons_base_path + current_theme;
+        theme_paths.append({default_theme_path, current_theme_path});
+        QIcon::setThemeName(current_theme);
    }

    QIcon::setThemeSearchPaths(theme_paths);
@@ -2998,6 +3001,10 @@ int main(int argc, char* argv[]) {
    setlocale(LC_ALL, "C");

    auto& system{Core::System::GetInstance()};
+
+    // Register Qt image interface
+    system.RegisterImageInterface(std::make_shared<QtImageInterface>());
+
    GMainWindow main_window(system);

    // Register frontend applets
@@ -3006,9 +3013,6 @@ int main(int argc, char* argv[]) {
    system.RegisterMiiSelector(std::make_shared<QtMiiSelector>(main_window));
    system.RegisterSoftwareKeyboard(std::make_shared<QtKeyboard>(main_window));

-    // Register Qt image interface
-    system.RegisterImageInterface(std::make_shared<QtImageInterface>());
-
 #ifdef __APPLE__
    // Register microphone permission check.
    system.RegisterMicPermissionCheck(&AppleAuthorization::CheckAuthorizationForMicrophone);
--- a/src/citra_qt/updater/updater.cpp
+++ b/src/citra_qt/updater/updater.cpp
@@ -17,7 +17,7 @@
 #include "common/file_util.h"
 #include "common/logging/log.h"

-#ifdef Q_OS_OSX
+#ifdef Q_OS_MACOS
 #define DEFAULT_TOOL_PATH QStringLiteral("../../../../maintenancetool")
 #else
 #define DEFAULT_TOOL_PATH QStringLiteral("../maintenancetool")
@@ -102,7 +102,7 @@ QString UpdaterPrivate::ToSystemExe(QString base_path) {
        return base_path + QStringLiteral(".exe");
    else
        return base_path;
-#elif defined(Q_OS_OSX)
+#elif defined(Q_OS_MACOS)
    if (base_path.endsWith(QStringLiteral(".app")))
        base_path.truncate(base_path.lastIndexOf(QStringLiteral(".")));
    return base_path + QStringLiteral(".app/Contents/MacOS/") + QFileInfo(base_path).fileName();
@@ -112,7 +112,7 @@ QString UpdaterPrivate::ToSystemExe(QString base_path) {
 }

 QFileInfo UpdaterPrivate::GetMaintenanceTool() const {
-#if defined(Q_OS_UNIX) && !defined(Q_OS_OSX)
+#if defined(Q_OS_UNIX) && !defined(Q_OS_MACOS)
    const auto appimage_path = QProcessEnvironment::systemEnvironment()
                                   .value(QStringLiteral("APPIMAGE"), {})
                                   .toStdString();
--- a/src/common/CMakeLists.txt
+++ b/src/common/CMakeLists.txt
@@ -124,6 +124,7 @@ add_library(citra_common STATIC
    serialization/boost_flat_set.h
    serialization/boost_small_vector.hpp
    serialization/boost_vector.hpp
+    static_lru_cache.h
    string_literal.h
    string_util.cpp
    string_util.h
@@ -174,7 +175,7 @@ endif()
 create_target_directory_groups(citra_common)

 target_link_libraries(citra_common PUBLIC fmt library-headers microprofile Boost::boost Boost::serialization Boost::iostreams)
-target_link_libraries(citra_common PRIVATE libzstd_static)
+target_link_libraries(citra_common PRIVATE zstd)

 if ("x86_64" IN_LIST ARCHITECTURE)
    target_link_libraries(citra_common PRIVATE xbyak)
--- a/src/common/file_util.cpp
+++ b/src/common/file_util.cpp
@@ -1155,6 +1155,43 @@ std::size_t IOFile::ReadImpl(void* data, std::size_t length, std::size_t data_si
    return std::fread(data, data_size, length, m_file);
 }

+#ifdef _WIN32
+static std::size_t pread(int fd, void* buf, size_t count, uint64_t offset) {
+    long unsigned int read_bytes = 0;
+    OVERLAPPED overlapped = {0};
+    HANDLE file = reinterpret_cast<HANDLE>(_get_osfhandle(fd));
+
+    overlapped.OffsetHigh = static_cast<uint32_t>(offset >> 32);
+    overlapped.Offset = static_cast<uint32_t>(offset & 0xFFFF'FFFFLL);
+    SetLastError(0);
+    bool ret = ReadFile(file, buf, static_cast<uint32_t>(count), &read_bytes, &overlapped);
+
+    if (!ret && GetLastError() != ERROR_HANDLE_EOF) {
+        errno = GetLastError();
+        return std::numeric_limits<std::size_t>::max();
+    }
+    return read_bytes;
+}
+#else
+#define pread ::pread
+#endif
+
+std::size_t IOFile::ReadAtImpl(void* data, std::size_t length, std::size_t data_size,
+                               std::size_t offset) {
+    if (!IsOpen()) {
+        m_good = false;
+        return std::numeric_limits<std::size_t>::max();
+    }
+
+    if (length == 0) {
+        return 0;
+    }
+
+    DEBUG_ASSERT(data != nullptr);
+
+    return pread(fileno(m_file), data, data_size * length, offset);
+}
+
 std::size_t IOFile::WriteImpl(const void* data, std::size_t length, std::size_t data_size) {
    if (!IsOpen()) {
        m_good = false;
--- a/src/common/file_util.h
+++ b/src/common/file_util.h
@@ -294,6 +294,18 @@ public:
        return items_read;
    }

+    template <typename T>
+    std::size_t ReadAtArray(T* data, std::size_t length, std::size_t offset) {
+        static_assert(std::is_trivially_copyable_v<T>,
+                      "Given array does not consist of trivially copyable objects");
+
+        std::size_t items_read = ReadAtImpl(data, length, sizeof(T), offset);
+        if (items_read != length)
+            m_good = false;
+
+        return items_read;
+    }
+
    template <typename T>
    std::size_t WriteArray(const T* data, std::size_t length) {
        static_assert(std::is_trivially_copyable_v<T>,
@@ -312,6 +324,12 @@ public:
        return ReadArray(reinterpret_cast<char*>(data), length);
    }

+    template <typename T>
+    std::size_t ReadAtBytes(T* data, std::size_t length, std::size_t offset) {
+        static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable");
+        return ReadAtArray(reinterpret_cast<char*>(data), length, offset);
+    }
+
    template <typename T>
    std::size_t WriteBytes(const T* data, std::size_t length) {
        static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable");
@@ -363,6 +381,8 @@ public:

 private:
    std::size_t ReadImpl(void* data, std::size_t length, std::size_t data_size);
+    std::size_t ReadAtImpl(void* data, std::size_t length, std::size_t data_size,
+                           std::size_t offset);
    std::size_t WriteImpl(const void* data, std::size_t length, std::size_t data_size);

    bool Open();
--- a/src/common/settings.cpp
+++ b/src/common/settings.cpp
@@ -46,8 +46,6 @@ std::string_view GetTextureFilterName(TextureFilter filter) {
        return "Anime4K";
    case TextureFilter::Bicubic:
        return "Bicubic";
-    case TextureFilter::NearestNeighbor:
-        return "NearestNeighbor";
    case TextureFilter::ScaleForce:
        return "ScaleForce";
    case TextureFilter::xBRZ:
@@ -59,6 +57,19 @@ std::string_view GetTextureFilterName(TextureFilter filter) {
    }
 }

+std::string_view GetTextureSamplingName(TextureSampling sampling) {
+    switch (sampling) {
+    case TextureSampling::GameControlled:
+        return "GameControlled";
+    case TextureSampling::NearestNeighbor:
+        return "NearestNeighbor";
+    case TextureSampling::Linear:
+        return "Linear";
+    default:
+        return "Invalid";
+    }
+}
+
 } // Anonymous namespace

 Values values = {};
@@ -87,6 +98,8 @@ void LogSettings() {
    log_setting("Renderer_PostProcessingShader", values.pp_shader_name.GetValue());
    log_setting("Renderer_FilterMode", values.filter_mode.GetValue());
    log_setting("Renderer_TextureFilter", GetTextureFilterName(values.texture_filter.GetValue()));
+    log_setting("Renderer_TextureSampling",
+                GetTextureSamplingName(values.texture_sampling.GetValue()));
    log_setting("Stereoscopy_Render3d", values.render_3d.GetValue());
    log_setting("Stereoscopy_Factor3d", values.factor_3d.GetValue());
    log_setting("Stereoscopy_MonoRenderOption", values.mono_render_option.GetValue());
@@ -175,6 +188,7 @@ void RestoreGlobalState(bool is_powered_on) {
    values.resolution_factor.SetGlobal(true);
    values.frame_limit.SetGlobal(true);
    values.texture_filter.SetGlobal(true);
+    values.texture_sampling.SetGlobal(true);
    values.layout_option.SetGlobal(true);
    values.swap_screen.SetGlobal(true);
    values.upright_screen.SetGlobal(true);
--- a/src/common/settings.h
+++ b/src/common/settings.h
@@ -72,10 +72,15 @@ enum class TextureFilter : u32 {
    None = 0,
    Anime4K = 1,
    Bicubic = 2,
-    NearestNeighbor = 3,
-    ScaleForce = 4,
-    xBRZ = 5,
-    MMPX = 6
+    ScaleForce = 3,
+    xBRZ = 4,
+    MMPX = 5,
+};
+
+enum class TextureSampling : u32 {
+    GameControlled = 0,
+    NearestNeighbor = 1,
+    Linear = 2,
 };

 namespace NativeButton {
@@ -451,6 +456,8 @@ struct Values {
    SwitchableSetting<u32, true> resolution_factor{1, 0, 10, "resolution_factor"};
    SwitchableSetting<u16, true> frame_limit{100, 0, 1000, "frame_limit"};
    SwitchableSetting<TextureFilter> texture_filter{TextureFilter::None, "texture_filter"};
+    SwitchableSetting<TextureSampling> texture_sampling{TextureSampling::GameControlled,
+                                                        "texture_sampling"};

    SwitchableSetting<LayoutOption> layout_option{LayoutOption::Default, "layout_option"};
    SwitchableSetting<bool> swap_screen{false, "swap_screen"};
--- a/src/common/static_lru_cache.h
+++ b/src/common/static_lru_cache.h
@@ -0,0 +1,113 @@
+// Modified version of: https://www.boost.org/doc/libs/1_79_0/boost/compute/detail/lru_cache.hpp
+// Most important change is the use of an array instead of a map, so that elements are
+// statically allocated. The insert and get methods have been merged into the request method.
+// Original license:
+//
+//---------------------------------------------------------------------------//
+// Copyright (c) 2013 Kyle Lutz <kyle.r.lutz@gmail.com>
+//
+// Distributed under the Boost Software License, Version 1.0
+// See accompanying file LICENSE_1_0.txt or copy at
+// http://www.boost.org/LICENSE_1_0.txt
+//
+// See http://boostorg.github.com/compute for more information.
+//---------------------------------------------------------------------------//
+#pragma once
+
+#include <array>
+#include <list>
+#include <tuple>
+#include <utility>
+
+namespace Common {
+
+// a cache which evicts the least recently used item when it is full
+// the cache elements are statically allocated.
+template <class Key, class Value, size_t Size>
+class StaticLRUCache {
+public:
+    using key_type = Key;
+    using value_type = Value;
+    using list_type = std::list<std::pair<Key, size_t>>;
+    using array_type = std::array<Value, Size>;
+
+    StaticLRUCache() = default;
+
+    ~StaticLRUCache() = default;
+
+    size_t size() const {
+        return m_list.size();
+    }
+
+    constexpr size_t capacity() const {
+        return m_array.size();
+    }
+
+    bool empty() const {
+        return m_list.empty();
+    }
+
+    bool contains(const key_type& key) const {
+        return find(key) != m_list.end();
+    }
+
+    // Requests an element from the cache. If it is not found,
+    // the element is inserted using its key.
+    // Returns whether the element was present in the cache
+    // and a reference to the element itself.
+    std::pair<bool, value_type&> request(const key_type& key) {
+        // lookup value in the cache
+        auto i = find(key);
+        if (i == m_list.cend()) {
+            size_t next_index = size();
+            // insert item into the cache, but first check if it is full
+            if (next_index >= capacity()) {
+                // cache is full, evict the least recently used item
+                next_index = evict();
+            }
+
+            // insert the new item
+            m_list.push_front(std::make_pair(key, next_index));
+            return std::pair<bool, value_type&>(false, m_array[next_index]);
+        }
+        // return the value, but first update its place in the most
+        // recently used list
+        if (i != m_list.cbegin()) {
+            // move item to the front of the most recently used list
+            auto backup = *i;
+            m_list.erase(i);
+            m_list.push_front(backup);
+
+            // return the value
+            return std::pair<bool, value_type&>(true, m_array[backup.second]);
+        } else {
+            // the item is already at the front of the most recently
+            // used list so just return it
+            return std::pair<bool, value_type&>(true, m_array[i->second]);
+        }
+    }
+
+    void clear() {
+        m_list.clear();
+    }
+
+private:
+    typename list_type::const_iterator find(const key_type& key) const {
+        return std::find_if(m_list.cbegin(), m_list.cend(),
+                            [&key](const auto& el) { return el.first == key; });
+    }
+
+    size_t evict() {
+        // evict item from the end of most recently used list
+        typename list_type::iterator i = --m_list.end();
+        size_t evicted_index = i->second;
+        m_list.erase(i);
+        return evicted_index;
+    }
+
+private:
+    array_type m_array;
+    list_type m_list;
+};
+
+} // namespace Common
--- a/src/common/zstd_compression.cpp
+++ b/src/common/zstd_compression.cpp
@@ -5,6 +5,7 @@
 #include <algorithm>
 #include <zstd.h>

+#include "common/logging/log.h"
 #include "common/zstd_compression.h"

 namespace Common::Compression {
--- a/src/core/file_sys/file_backend.h
+++ b/src/core/file_sys/file_backend.h
@@ -86,6 +86,20 @@ public:
     */
    virtual void Flush() const = 0;

+    /**
+     * Whether the backend supports cached reads.
+     */
+    virtual bool AllowsCachedReads() const {
+        return false;
+    }
+
+    /**
+     * Whether the cache is ready for a specified offset and length.
+     */
+    virtual bool CacheReady(std::size_t file_offset, std::size_t length) {
+        return false;
+    }
+
 protected:
    std::unique_ptr<DelayGenerator> delay_generator;

--- a/src/core/file_sys/ivfc_archive.h
+++ b/src/core/file_sys/ivfc_archive.h
@@ -131,6 +131,14 @@ public:
    }
    void Flush() const override {}

+    bool AllowsCachedReads() const override {
+        return romfs_file->AllowsCachedReads();
+    }
+
+    bool CacheReady(std::size_t file_offset, std::size_t length) override {
+        return romfs_file->CacheReady(file_offset, length);
+    }
+
 private:
    std::shared_ptr<RomFSReader> romfs_file;

--- a/src/core/file_sys/layered_fs.h
+++ b/src/core/file_sys/layered_fs.h
@@ -53,6 +53,14 @@ public:

    bool DumpRomFS(const std::string& target_path);

+    bool AllowsCachedReads() const override {
+        return false;
+    }
+
+    bool CacheReady(std::size_t file_offset, std::size_t length) override {
+        return false;
+    }
+
 private:
    struct File;
    struct Directory {
--- a/src/core/file_sys/romfs_reader.cpp
+++ b/src/core/file_sys/romfs_reader.cpp
@@ -1,4 +1,5 @@
 #include <algorithm>
+#include <vector>
 #include <cryptopp/aes.h>
 #include <cryptopp/modes.h>
 #include "common/archives.h"
@@ -9,17 +10,102 @@ SERIALIZE_EXPORT_IMPL(FileSys::DirectRomFSReader)
 namespace FileSys {

 std::size_t DirectRomFSReader::ReadFile(std::size_t offset, std::size_t length, u8* buffer) {
+    length = std::min(length, static_cast<std::size_t>(data_size) - offset);
    if (length == 0)
        return 0; // Crypto++ does not like zero size buffer
-    file.Seek(file_offset + offset, SEEK_SET);
-    std::size_t read_length = std::min(length, static_cast<std::size_t>(data_size) - offset);
-    read_length = file.ReadBytes(buffer, read_length);
-    if (is_encrypted) {
-        CryptoPP::CTR_Mode<CryptoPP::AES>::Decryption d(key.data(), key.size(), ctr.data());
-        d.Seek(crypto_offset + offset);
-        d.ProcessData(buffer, buffer, read_length);
+
+    const auto segments = BreakupRead(offset, length);
+    size_t read_progress = 0;
+
+    // Skip cache if the read is too big
+    if (segments.size() == 1 && segments[0].second > cache_line_size) {
+        length = file.ReadAtBytes(buffer, length, file_offset + offset);
+        if (is_encrypted) {
+            CryptoPP::CTR_Mode<CryptoPP::AES>::Decryption d(key.data(), key.size(), ctr.data());
+            d.Seek(crypto_offset + offset);
+            d.ProcessData(buffer, buffer, length);
+        }
+        // LOG_INFO(Service_FS, "Cache SKIP: offset={}, length={}", offset, length);
+        return length;
    }
-    return read_length;
+
+    // TODO(PabloMK7): Make cache thread safe, read the comment in CacheReady function.
+    // std::unique_lock<std::shared_mutex> read_guard(cache_mutex);
+    for (const auto& seg : segments) {
+        size_t read_size = cache_line_size;
+        size_t page = OffsetToPage(seg.first);
+        // Check if segment is in cache
+        auto cache_entry = cache.request(page);
+        if (!cache_entry.first) {
+            // If not found, read from disk and cache the data
+            read_size = file.ReadAtBytes(cache_entry.second.data(), read_size, file_offset + page);
+            if (is_encrypted && read_size) {
+                CryptoPP::CTR_Mode<CryptoPP::AES>::Decryption d(key.data(), key.size(), ctr.data());
+                d.Seek(crypto_offset + page);
+                d.ProcessData(cache_entry.second.data(), cache_entry.second.data(), read_size);
+            }
+            // LOG_INFO(Service_FS, "Cache MISS: page={}, length={}, into={}", page, seg.second,
+            //          (seg.first - page));
+        } else {
+            // LOG_INFO(Service_FS, "Cache HIT: page={}, length={}, into={}", page, seg.second,
+            //          (seg.first - page));
+        }
+        size_t copy_amount =
+            (read_size > (seg.first - page))
+                ? std::min((seg.first - page) + seg.second, read_size) - (seg.first - page)
+                : 0;
+        std::memcpy(buffer + read_progress, cache_entry.second.data() + (seg.first - page),
+                    copy_amount);
+        read_progress += copy_amount;
+    }
+    return read_progress;
+}
+
+bool DirectRomFSReader::AllowsCachedReads() const {
+    return true;
+}
+
+bool DirectRomFSReader::CacheReady(std::size_t file_offset, std::size_t length) {
+    auto segments = BreakupRead(file_offset, length);
+    if (segments.size() == 1 && segments[0].second > cache_line_size) {
+        return false;
+    } else {
+        // TODO(PabloMK7): Since the LRU cache is not thread safe, a lock must be used.
+        // However, this completely breaks the point of using a cache, because
+        // smaller reads may be blocked by bigger reads. For now, always return
+        // data being in cache to prevent the need of a lock, and only read data
+        // asynchronously if it is too big to use the cache.
+        /*
+        std::shared_lock<std::shared_mutex> read_guard(cache_mutex);
+        for (auto it = segments.begin(); it != segments.end(); it++) {
+            if (!cache.contains(OffsetToPage(it->first)))
+                return false;
+        }
+        */
+        return true;
+    }
+}
+
+std::vector<std::pair<std::size_t, std::size_t>> DirectRomFSReader::BreakupRead(
+    std::size_t offset, std::size_t length) {
+
+    std::vector<std::pair<std::size_t, std::size_t>> ret;
+
+    // Reads bigger than the cache line size will probably never hit again
+    if (length > cache_line_size) {
+        ret.push_back(std::make_pair(offset, length));
+        return ret;
+    }
+
+    size_t curr_offset = offset;
+    while (length) {
+        size_t next_page = OffsetToPage(curr_offset + cache_line_size);
+        size_t curr_page_len = std::min(length, next_page - curr_offset);
+        ret.push_back(std::make_pair(curr_offset, curr_page_len));
+        curr_offset = next_page;
+        length -= curr_page_len;
+    }
+    return ret;
 }

 } // namespace FileSys
--- a/src/core/file_sys/romfs_reader.h
+++ b/src/core/file_sys/romfs_reader.h
@@ -1,11 +1,14 @@
 #pragma once

 #include <array>
+#include <shared_mutex>
 #include <boost/serialization/array.hpp>
 #include <boost/serialization/base_object.hpp>
 #include <boost/serialization/export.hpp>
+#include "common/alignment.h"
 #include "common/common_types.h"
 #include "common/file_util.h"
+#include "common/static_lru_cache.h"

 namespace FileSys {

@@ -18,6 +21,8 @@ public:

    virtual std::size_t GetSize() const = 0;
    virtual std::size_t ReadFile(std::size_t offset, std::size_t length, u8* buffer) = 0;
+    virtual bool AllowsCachedReads() const = 0;
+    virtual bool CacheReady(std::size_t file_offset, std::size_t length) = 0;

 private:
    template <class Archive>
@@ -48,6 +53,10 @@ public:

    std::size_t ReadFile(std::size_t offset, std::size_t length, u8* buffer) override;

+    bool AllowsCachedReads() const override;
+
+    bool CacheReady(std::size_t file_offset, std::size_t length) override;
+
 private:
    bool is_encrypted;
    FileUtil::IOFile file;
@@ -57,8 +66,23 @@ private:
    u64 crypto_offset;
    u64 data_size;

+    // Total cache size: 128KB
+    static constexpr size_t cache_line_size = (1 << 13); // About 8KB
+    static constexpr size_t cache_line_count = 16;
+
+    Common::StaticLRUCache<std::size_t, std::array<u8, cache_line_size>, cache_line_count> cache;
+    // TODO(PabloMK7): Make cache thread safe, read the comment in CacheReady function.
+    // std::shared_mutex cache_mutex;
+
    DirectRomFSReader() = default;

+    std::size_t OffsetToPage(std::size_t offset) {
+        return Common::AlignDown<std::size_t>(offset, cache_line_size);
+    }
+
+    std::vector<std::pair<std::size_t, std::size_t>> BreakupRead(std::size_t offset,
+                                                                 std::size_t length);
+
    template <class Archive>
    void serialize(Archive& ar, const unsigned int) {
        ar& boost::serialization::base_object<RomFSReader>(*this);
--- a/src/core/hle/kernel/address_arbiter.cpp
+++ b/src/core/hle/kernel/address_arbiter.cpp
@@ -6,6 +6,7 @@
 #include "common/archives.h"
 #include "common/common_types.h"
 #include "common/logging/log.h"
+#include "core/global.h"
 #include "core/hle/kernel/address_arbiter.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/kernel.h"
@@ -20,7 +21,7 @@ void AddressArbiter::WaitThread(std::shared_ptr<Thread> thread, VAddr wait_addre
    waiting_threads.emplace_back(std::move(thread));
 }

-u64 AddressArbiter::ResumeAllThreads(VAddr address) {
+void AddressArbiter::ResumeAllThreads(VAddr address) {
    // Determine which threads are waiting on this address, those should be woken up.
    auto itr = std::stable_partition(waiting_threads.begin(), waiting_threads.end(),
                                     [address](const auto& thread) {
@@ -30,15 +31,13 @@ u64 AddressArbiter::ResumeAllThreads(VAddr address) {
                                     });

    // Wake up all the found threads
-    const u64 num_threads = std::distance(itr, waiting_threads.end());
    std::for_each(itr, waiting_threads.end(), [](auto& thread) { thread->ResumeFromWait(); });

    // Remove the woken up threads from the wait list.
    waiting_threads.erase(itr, waiting_threads.end());
-    return num_threads;
 }

-bool AddressArbiter::ResumeHighestPriorityThread(VAddr address) {
+std::shared_ptr<Thread> AddressArbiter::ResumeHighestPriorityThread(VAddr address) {
    // Determine which threads are waiting on this address, those should be considered for wakeup.
    auto matches_start = std::stable_partition(
        waiting_threads.begin(), waiting_threads.end(), [address](const auto& thread) {
@@ -55,15 +54,14 @@ bool AddressArbiter::ResumeHighestPriorityThread(VAddr address) {
                                    return lhs->current_priority < rhs->current_priority;
                                });

-    if (itr == waiting_threads.end()) {
-        return false;
-    }
+    if (itr == waiting_threads.end())
+        return nullptr;

    auto thread = *itr;
    thread->ResumeFromWait();
-    waiting_threads.erase(itr);

-    return true;
+    waiting_threads.erase(itr);
+    return thread;
 }

 AddressArbiter::AddressArbiter(KernelSystem& kernel)
@@ -109,28 +107,17 @@ ResultCode AddressArbiter::ArbitrateAddress(std::shared_ptr<Thread> thread, Arbi
    switch (type) {

    // Signal thread(s) waiting for arbitrate address...
-    case ArbitrationType::Signal: {
-        u64 num_threads{};
-
+    case ArbitrationType::Signal:
        // Negative value means resume all threads
        if (value < 0) {
-            num_threads = ResumeAllThreads(address);
+            ResumeAllThreads(address);
        } else {
            // Resume first N threads
-            for (s32 i = 0; i < value; i++) {
-                num_threads += ResumeHighestPriorityThread(address);
-            }
-        }
-
-        // Prevents lag from low priority threads that spam svcArbitrateAddress and wake no threads
-        // The tick count is taken directly from official HOS kernel. The priority value is one less
-        // than official kernel as the affected FMV threads dont meet the priority threshold of 50.
-        // TODO: Revisit this when scheduler is rewritten and adjust if there isn't a problem there.
-        if (num_threads == 0 && thread->current_priority >= 49) {
-            kernel.current_cpu->GetTimer().AddTicks(1614u);
+            for (int i = 0; i < value; i++)
+                ResumeHighestPriorityThread(address);
        }
        break;
-    }
+
    // Wait current thread (acquire the arbiter)...
    case ArbitrationType::WaitIfLessThan:
        if ((s32)kernel.memory.Read32(address) < value) {
--- a/src/core/hle/kernel/address_arbiter.h
+++ b/src/core/hle/kernel/address_arbiter.h
@@ -65,11 +65,11 @@ private:
    void WaitThread(std::shared_ptr<Thread> thread, VAddr wait_address);

    /// Resume all threads found to be waiting on the address under this address arbiter
-    u64 ResumeAllThreads(VAddr address);
+    void ResumeAllThreads(VAddr address);

    /// Resume one thread found to be waiting on the address under this address arbiter and return
    /// the resumed thread.
-    bool ResumeHighestPriorityThread(VAddr address);
+    std::shared_ptr<Thread> ResumeHighestPriorityThread(VAddr address);

    /// Threads waiting for the address arbiter to be signaled.
    std::vector<std::shared_ptr<Thread>> waiting_threads;
--- a/src/core/hle/kernel/hle_ipc.h
+++ b/src/core/hle/kernel/hle_ipc.h
@@ -221,6 +221,13 @@ public:
        return session;
    }

+    /**
+     * Returns the client thread that made the service request.
+     */
+    std::shared_ptr<Thread> ClientThread() const {
+        return thread;
+    }
+
    class WakeupCallback {
    public:
        virtual ~WakeupCallback() = default;
--- a/src/core/hle/kernel/process.cpp
+++ b/src/core/hle/kernel/process.cpp
@@ -1,7 +1,7 @@
 // Copyright 2015 Citra Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
-#include "core/hle/result.h"
+
 #include <algorithm>
 #include <memory>
 #include <boost/serialization/array.hpp>
@@ -459,7 +459,8 @@ ResultVal<VAddr> Process::AllocateThreadLocalStorage() {
    return tls_address;
 }

-ResultCode Process::Map(VAddr target, VAddr source, u32 size, VMAPermission perms, bool privileged) {
+ResultCode Process::Map(VAddr target, VAddr source, u32 size, VMAPermission perms,
+                        bool privileged) {
    LOG_DEBUG(Kernel, "Map memory target={:08X}, source={:08X}, size={:08X}, perms={:08X}", target,
              source, size, perms);
    if (!privileged && (source < Memory::HEAP_VADDR || source + size > Memory::HEAP_VADDR_END ||
@@ -501,10 +502,8 @@ ResultCode Process::Map(VAddr target, VAddr source, u32 size, VMAPermission perm
    CASCADE_CODE(vm_manager.ChangeMemoryState(source, size, MemoryState::Private,
                                              VMAPermission::ReadWrite, source_state, source_perm));

-    // Retrieve backing blocks from the source address space
    CASCADE_RESULT(auto backing_blocks, vm_manager.GetBackingBlocksForRange(source, size));
    VAddr interval_target = target;
-
    for (const auto& backing_block : backing_blocks) {
        auto backing_memory = kernel.memory.GetFCRAMRef(backing_block.lower());
        auto block_size = backing_block.upper() - backing_block.lower();
@@ -517,36 +516,6 @@ ResultCode Process::Map(VAddr target, VAddr source, u32 size, VMAPermission perm

    return RESULT_SUCCESS;
 }
-
-ResultCode Process::MapInterprocess(Process& src, VAddr target, VAddr source, u32 size, VMAPermission perms) {
-    u32 cursor = 0;
-    while (true) {
-        const auto source_iter = src.vm_manager.FindVMA(source + cursor);
-        auto& source_vma = source_iter->second;
-
-        // Skip non-allocated blocks
-        if (source_vma.type == VMAType::Free) {
-            cursor += source_vma.size;
-            continue;
-        }
-
-        ASSERT(source_vma.type == VMAType::BackingMemory);
-        ASSERT(source_vma.backing_memory);
-        ASSERT(source_vma.base == source + cursor);
-
-        if (source_vma.size + cursor > size) {
-            break;
-        }
-
-        const auto target_vma = vm_manager.MapBackingMemory(target + cursor, source_vma.backing_memory,
-                                                            source_vma.size, MemoryState::Shared);
-        ASSERT(target_vma.Succeeded());
-        cursor += source_vma.size;
-    }
-
-    return RESULT_SUCCESS;
-}
-
 ResultCode Process::Unmap(VAddr target, VAddr source, u32 size, VMAPermission perms,
                          bool privileged) {
    LOG_DEBUG(Kernel, "Unmap memory target={:08X}, source={:08X}, size={:08X}, perms={:08X}",
--- a/src/core/hle/kernel/process.h
+++ b/src/core/hle/kernel/process.h
@@ -209,6 +209,9 @@ public:
     */
    void Exit();

+    ///////////////////////////////////////////////////////////////////////////////////////////////
+    // Memory Management
+
    VMManager vm_manager;

    u32 memory_used = 0;
@@ -240,8 +243,6 @@ public:

    ResultCode Map(VAddr target, VAddr source, u32 size, VMAPermission perms,
                   bool privileged = false);
-    ResultCode MapInterprocess(Process& src, VAddr target, VAddr source,
-                               u32 size, VMAPermission perms);
    ResultCode Unmap(VAddr target, VAddr source, u32 size, VMAPermission perms,
                     bool privileged = false);

--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -1,16 +1,14 @@
 // Copyright 2014 Citra Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
-#pragma optimize("", off)
+
 #include <algorithm>
 #include <array>
 #include <fmt/format.h>
-#include "common/alignment.h"
 #include "common/archives.h"
 #include "common/logging/log.h"
 #include "common/microprofile.h"
 #include "common/scm_rev.h"
-#include "common/scope_exit.h"
 #include "core/arm/arm_interface.h"
 #include "core/core.h"
 #include "core/core_timing.h"
@@ -440,10 +438,6 @@ private:
    ResultCode UnmapProcessMemoryEx(Handle process, u32 dst_address, u32 size);
    ResultCode ControlProcess(Handle process_handle, u32 process_OP, u32 varg2, u32 varg3);

-    ResultCode ControlProcessMemory(Handle process, u32 addr0, u32 addr1, u32 size, u32 type, u32 perm);
-    ResultCode MapProcessMemory(Handle process, u32 dest_address, u32 size);
-    ResultCode UnmapProcessMemory(Handle process, u32 start_address, u32 size);
-
    struct FunctionDef {
        using Func = void (SVC::*)();

@@ -2109,283 +2103,6 @@ ResultCode SVC::ControlProcess(Handle process_handle, u32 process_OP, u32 varg2,
    }
 }

-ResultCode SVC::ControlProcessMemory(Handle process, u32 addr0, u32 addr1, u32 size, u32 operation, u32 permissions) {
-    LOG_DEBUG(Kernel_SVC,
-              "called operation=0x{:08X}, addr0=0x{:08X}, addr1=0x{:08X}, "
-              "size=0x{:X}, permissions=0x{:08X}",
-              operation, addr0, addr1, size, permissions);
-
-    if (!addr0) {
-        return ResultCode{0xD8E007F6};
-    }
-
-    if ((addr0 & Memory::CITRA_PAGE_MASK) != 0 || (addr1 & Memory::CITRA_PAGE_MASK) != 0) {
-        return ERR_MISALIGNED_ADDRESS;
-    }
-    if ((size & Memory::CITRA_PAGE_MASK) != 0) {
-        return ERR_MISALIGNED_SIZE;
-    }
-
-    u32 region = operation & MEMOP_REGION_MASK;
-    operation &= ~MEMOP_REGION_MASK;
-
-    if (region != 0) {
-        LOG_WARNING(Kernel_SVC, "ControlMemory with specified region not supported, region={:X}",
-                    region);
-    }
-
-    VMAPermission vma_permissions = (VMAPermission)permissions;
-
-    const auto dst_process = kernel.GetCurrentProcess()->handle_table.Get<Process>(process);
-
-    SCOPE_EXIT({ system.GetRunningCore().ClearInstructionCache(); });
-
-    switch (operation & MEMOP_OPERATION_MASK) {
-    case MEMOP_MAP: {
-        CASCADE_CODE(dst_process->Map(addr0, addr1, size, vma_permissions, true));
-        break;
-    }
-    case MEMOP_UNMAP: {
-        CASCADE_CODE(dst_process->Unmap(addr0, addr1, size, vma_permissions, true));
-        break;
-    }
-    case MEMOP_PROTECT: {
-        CASCADE_CODE(dst_process->vm_manager.ReprotectRange(addr0, size, vma_permissions));
-        break;
-    }
-    default:
-        LOG_ERROR(Kernel_SVC, "unknown operation=0x%08X", operation);
-        return ResultCode(ErrorDescription::NotImplemented, ErrorModule::Kernel,
-                          ErrorSummary::NotSupported, ErrorLevel::Fatal);
-    }
-
-    return RESULT_SUCCESS;
-}
-
-ResultCode SVC::MapProcessMemory(Handle process, u32 start_addr, u32 size) {
-    const auto dst_process = kernel.GetCurrentProcess();
-    const auto src_process = kernel.GetCurrentProcess()->handle_table.Get<Process>(process);
-    if (!dst_process) {
-        return ERR_INVALID_HANDLE;
-    }
-
-           // Map up to 0x3F00000 bytes starting from 0x100000 from the source process into the current
-           // process' address space at the specified address. Note: Non-page aligned sizes do not return
-           // an error, the lower bits are just discarded.
-    const u32 code_size =
-        Common::AlignDown(std::min<u32>(size, Memory::PROCESS_IMAGE_MAX_SIZE), Memory::CITRA_PAGE_SIZE);
-    ResultCode result = dst_process->MapInterprocess(*src_process, start_addr, Memory::PROCESS_IMAGE_VADDR,
-                                                     code_size, VMAPermission::ReadWrite);
-    ASSERT(result.IsSuccess());
-
-    static constexpr u32 MaxHeapSize = 0x6000000;
-    static constexpr u32 HeapOffset = 0x7F00000;
-
-           // Now map up to 0x6000000 bytes starting from 0x8000000 from the source process into the
-           // current process' address space at the specified address + 0x7F00000.
-    const u32 heap_size =
-        Common::AlignDown(std::min<u32>(size - HeapOffset, MaxHeapSize), Memory::CITRA_PAGE_SIZE);
-    if (heap_size > 0) {
-        result = dst_process->MapInterprocess(*src_process, start_addr + HeapOffset, Memory::HEAP_VADDR,
-                                              heap_size, VMAPermission::ReadWrite);
-        ASSERT(result.IsSuccess());
-    }
-
-    return RESULT_SUCCESS;
-}
-
-ResultCode SVC::UnmapProcessMemory(Handle process, VAddr start_addr, u32 size) {
-    const u32 code_size =
-        Common::AlignDown(std::min<u32>(size, Memory::PROCESS_IMAGE_MAX_SIZE), Memory::CITRA_PAGE_SIZE);
-
-    ResultCode result = kernel.GetCurrentProcess()->vm_manager.UnmapRange(start_addr, code_size);
-    ASSERT(result.IsSuccess());
-
-    static constexpr u32 MaxHeapSize = 0x6000000;
-    static constexpr u32 HeapOffset = 0x7F00000;
-
-    const u32 heap_size =
-        Common::AlignDown(std::min<u32>(size - HeapOffset, MaxHeapSize), Memory::CITRA_PAGE_SIZE);
-    result = kernel.GetCurrentProcess()->vm_manager.UnmapRange(start_addr + HeapOffset, heap_size);
-    ASSERT(result.IsSuccess());
-
-    system.GetRunningCore().ClearInstructionCache();
-    return RESULT_SUCCESS;
-}
-
-/*ResultCode SVC::ControlProcessMemory(Handle process, u32 addr0, u32 addr1, u32 size, u32 operation, u32 permissions) {
-    LOG_DEBUG(Kernel_SVC,
-              "called operation=0x{:08X}, addr0=0x{:08X}, addr1=0x{:08X}, "
-              "size=0x{:X}, permissions=0x{:08X}",
-              operation, addr0, addr1, size, permissions);
-
-    if (!addr0) {
-        return ResultCode{0xD8E007F6};
-    }
-
-    if ((addr0 & Memory::CITRA_PAGE_MASK) != 0 || (addr1 & Memory::CITRA_PAGE_MASK) != 0) {
-        return ERR_MISALIGNED_ADDRESS;
-    }
-    if ((size & Memory::CITRA_PAGE_MASK) != 0) {
-        return ERR_MISALIGNED_SIZE;
-    }
-
-    u32 region = operation & MEMOP_REGION_MASK;
-    operation &= ~MEMOP_REGION_MASK;
-
-    if (region != 0) {
-        LOG_WARNING(Kernel_SVC, "ControlMemory with specified region not supported, region={:X}",
-                    region);
-    }
-
-    VMAPermission vma_permissions = (VMAPermission)permissions;
-
-    const auto dst_process = kernel.GetCurrentProcess()->handle_table.Get<Process>(process);
-
-    SCOPE_EXIT({ system.GetRunningCore().ClearInstructionCache(); });
-
-    switch (operation) {
-    case MEMOP_MAP: {
-
-        if (addr0 != addr1) {
-            // TODO(Subv): Check if the addresses are valid.
-            auto addr1_vma = dst_process->vm_manager.FindVMA(addr1);
-            ASSERT(addr1_vma->second.meminfo_state == MemoryState::Private);
-            ASSERT(addr1 + size <= addr1_vma->second.base + addr1_vma->second.size);
-            ASSERT((static_cast<u32>(addr1_vma->second.permissions) &
-                    static_cast<u32>(VMAPermission::ReadWrite)) ==
-                   static_cast<u32>(VMAPermission::ReadWrite));
-
-            auto addr0_vma = dst_process->vm_manager.FindVMA(addr0);
-            ASSERT(addr0_vma->second.meminfo_state == MemoryState::Free);
-            ASSERT(addr0 + size <= addr0_vma->second.base + addr0_vma->second.size);
-
-            // Note: The real kernel will crash if this fails and the addresses are valid.
-            auto result = dst_process->vm_manager.AliasMemory(
-                addr0, addr1, size, MemoryState::AliasCode, MemoryState::Locked);
-            ASSERT(result.Succeeded());
-            ASSERT(dst_process->vm_manager.ReprotectRange(addr0, size, vma_permissions).IsSuccess());
-            ASSERT(
-                dst_process->vm_manager.ReprotectRange(addr1, size, VMAPermission::None).IsSuccess());
-        } else {
-            // This code path acts like PROTECT but also changes the memory state.
-            CASCADE_CODE(dst_process->vm_manager.ChangeMemoryState(
-                addr0, size, MemoryState::Private, VMAPermission::ReadWrite, MemoryState::AliasCode,
-                vma_permissions));
-        }
-        break;
-    }
-
-    case MEMOP_UNMAP: {
-        if (addr0 != addr1) {
-            // TODO(Subv): Check if the addresses are valid.
-            auto addr1_vma = dst_process->vm_manager.FindVMA(addr1);
-            if (addr1_vma->second.meminfo_state != MemoryState::Locked) {
-                return ResultCode(0xE0A01BF5);
-            }
-            ASSERT(addr1 + size <= addr1_vma->second.base + addr1_vma->second.size);
-
-            auto addr0_vma = dst_process->vm_manager.FindVMA(addr0);
-            ASSERT(addr0_vma->second.meminfo_state == MemoryState::AliasCode);
-            ASSERT(addr0 + size <= addr0_vma->second.base + addr0_vma->second.size);
-
-            ResultCode result = dst_process->vm_manager.UnmapRange(addr0, size);
-            ASSERT(result.IsSuccess());
-
-            result = dst_process->vm_manager.ChangeMemoryState(addr1, size, MemoryState::Locked,
-                                                           VMAPermission::None,
-                                                           MemoryState::Private, vma_permissions);
-            ASSERT(result.IsSuccess());
-            return result;
-        } else {
-            // This code path acts like PROTECT but also changes the memory state.
-            CASCADE_CODE(dst_process->vm_manager.ChangeMemoryState(
-                addr0, size, MemoryState::AliasCode, VMAPermission::None, MemoryState::Private,
-                vma_permissions));
-        }
-        break;
-    }
-
-    case MEMOP_PROTECT: {
-        ResultCode result = dst_process->vm_manager.ReprotectRange(addr0, size, vma_permissions);
-        if (result.IsError())
-            return result;
-        break;
-    }
-
-    default:
-        LOG_ERROR(Kernel_SVC, "unknown operation=0x%08X", operation);
-        return ResultCode(ErrorDescription::NotImplemented, ErrorModule::Kernel,
-                          ErrorSummary::NotSupported, ErrorLevel::Fatal);
-    }
-
-    return RESULT_SUCCESS;
-}
-
-ResultCode SVC::MapProcessMemory(Handle process, u32 start_addr, u32 size) {
-    const auto dst_process = kernel.GetCurrentProcess();
-    const auto src_process = kernel.GetCurrentProcess()->handle_table.Get<Process>(process);
-    if (!dst_process) {
-        return ERR_INVALID_HANDLE;
-    }
-
-    SCOPE_EXIT({ system.GetRunningCore().ClearInstructionCache(); });
-
-    // Map up to 0x3F00000 bytes starting from 0x100000 from the source process into the current
-    // process' address space at the specified address. Note: Non-page aligned sizes do not return
-    // an error, the lower bits are just discarded.
-    u32 code_size =
-        Common::AlignDown(std::min<u32>(size, Memory::PROCESS_IMAGE_MAX_SIZE), Memory::CITRA_PAGE_SIZE);
-
-    auto vma_handle = dst_process->vm_manager.FindVMA(start_addr);
-    if (vma_handle->second.type != Kernel::VMAType::Free) {
-        return ResultCode(ErrorDescription::InvalidAddress, ErrorModule::OS,
-                          ErrorSummary::InvalidState, ErrorLevel::Usage);
-    }
-
-    auto result = dst_process->vm_manager.AliasMemory(
-        start_addr, Memory::PROCESS_IMAGE_VADDR, code_size, Kernel::MemoryState::Shared,
-        Kernel::VMAPermission::ReadWrite, src_process->vm_manager);
-    // TODO(Subv): Handle error cases.
-    ASSERT(result.IsSuccess());
-
-   // Now map up to 0x6000000 bytes starting from 0x8000000 from the source process into the
-   // current process' address space at the specified address + 0x7F00000.
-    static constexpr u32 MaxHeapSize = 0x6000000;
-    static constexpr u32 HeapOffset = 0x7F00000;
-    u32 heap_size =
-        Common::AlignDown(std::min<u32>(size - HeapOffset, MaxHeapSize), Memory::CITRA_PAGE_SIZE);
-    if (heap_size > 0) {
-        result = dst_process->vm_manager.AliasMemory(
-            start_addr + HeapOffset, Memory::HEAP_VADDR, heap_size, Kernel::MemoryState::Shared,
-            Kernel::VMAPermission::ReadWrite, src_process->vm_manager);
-        // TODO(Subv): Handle error cases.
-        ASSERT(result.IsSuccess());
-    }
-    return RESULT_SUCCESS;
-}
-
-ResultCode SVC::UnmapProcessMemory(Handle process, VAddr start_addr, u32 size) {
-    const u32 code_size =
-        Common::AlignDown(std::min<u32>(size, Memory::PROCESS_IMAGE_MAX_SIZE), Memory::CITRA_PAGE_SIZE);
-
-    SCOPE_EXIT({ system.GetRunningCore().ClearInstructionCache(); });
-
-    ResultCode result = kernel.GetCurrentProcess()->vm_manager.UnmapRange(start_addr, code_size);
-    ASSERT(result.IsSuccess());
-
-    static constexpr u32 MaxHeapSize = 0x6000000;
-    static constexpr u32 HeapOffset = 0x7F00000;
-
-    const u32 heap_size =
-        Common::AlignDown(std::min<u32>(size - HeapOffset, MaxHeapSize), Memory::CITRA_PAGE_SIZE);
-    result = kernel.GetCurrentProcess()->vm_manager.UnmapRange(start_addr + HeapOffset, heap_size);
-    ASSERT(result.IsSuccess());
-
-    system.GetRunningCore().ClearInstructionCache();
-    return RESULT_SUCCESS;
-}*/
-
 const std::array<SVC::FunctionDef, 180> SVC::SVC_Table{{
    {0x00, nullptr, "Unknown"},
    {0x01, &SVC::Wrap<&SVC::ControlMemory>, "ControlMemory"},
@@ -2499,9 +2216,9 @@ const std::array<SVC::FunctionDef, 180> SVC::SVC_Table{{
    {0x6D, nullptr, "GetDebugThreadParam"},
    {0x6E, nullptr, "Unknown"},
    {0x6F, nullptr, "Unknown"},
-    {0x70, &SVC::Wrap<&SVC::ControlProcessMemory>, "ControlProcessMemory"},
-    {0x71, &SVC::Wrap<&SVC::MapProcessMemory>, "MapProcessMemory"},
-    {0x72, &SVC::Wrap<&SVC::UnmapProcessMemory>, "UnmapProcessMemory"},
+    {0x70, nullptr, "ControlProcessMemory"},
+    {0x71, nullptr, "MapProcessMemory"},
+    {0x72, nullptr, "UnmapProcessMemory"},
    {0x73, nullptr, "CreateCodeSet"},
    {0x74, nullptr, "RandomStub"},
    {0x75, nullptr, "CreateProcess"},
--- a/src/core/hle/kernel/vm_manager.cpp
+++ b/src/core/hle/kernel/vm_manager.cpp
@@ -1,6 +1,7 @@
 // Copyright 2015 Citra Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+
 #include <algorithm>
 #include <iterator>
 #include "common/assert.h"
@@ -306,11 +307,11 @@ ResultVal<VMManager::VMAIter> VMManager::CarveVMARange(VAddr target, u32 size) {
    ASSERT(size > 0);

    VMAIter begin_vma = StripIterConstness(FindVMA(target));
-    //const VMAIter i_end = vma_map.lower_bound(target_end);
-    //if (std::any_of(begin_vma, i_end,
-    //                [](const auto& entry) { return entry.second.type == VMAType::Free; })) {
-    //    return ERR_INVALID_ADDRESS_STATE;
-    //}
+    const VMAIter i_end = vma_map.lower_bound(target_end);
+    if (std::any_of(begin_vma, i_end,
+                    [](const auto& entry) { return entry.second.type == VMAType::Free; })) {
+        return ERR_INVALID_ADDRESS_STATE;
+    }

    if (target != begin_vma->second.base) {
        begin_vma = SplitVMA(begin_vma, target - begin_vma->second.base);
@@ -341,7 +342,6 @@ VMManager::VMAIter VMManager::SplitVMA(VMAIter vma_handle, u32 offset_in_vma) {
    case VMAType::Free:
        break;
    case VMAType::BackingMemory:
-    case VMAType::MemoryAlias:
        new_vma.backing_memory += offset_in_vma;
        break;
    case VMAType::MMIO:
@@ -379,7 +379,6 @@ void VMManager::UpdatePageTableForVMA(const VirtualMemoryArea& vma) {
        memory.UnmapRegion(*page_table, vma.base, vma.size);
        break;
    case VMAType::BackingMemory:
-    case VMAType::MemoryAlias:
        memory.MapMemoryRegion(*page_table, vma.base, vma.size, vma.backing_memory);
        break;
    case VMAType::MMIO:
@@ -397,97 +396,21 @@ ResultVal<MemoryRegionInfo::IntervalSet> VMManager::GetBackingBlocksForRange(VAd
    MemoryRegionInfo::IntervalSet backing_blocks;
    VAddr interval_target = address;
    while (interval_target != address + size) {
-        const auto it = FindVMA(interval_target);
-        if (it == vma_map.end() || (it->second.type != VMAType::BackingMemory && it->second.type != VMAType::Free)) {
+        auto vma = FindVMA(interval_target);
+        if (vma->second.type != VMAType::BackingMemory) {
            LOG_ERROR(Kernel, "Trying to use already freed memory");
            return ERR_INVALID_ADDRESS_STATE;
        }

-        const auto vma = it->second;
-        const VAddr interval_end = std::min(address + size, vma.base + vma.size);
-        const u32 interval_size = interval_end - interval_target;
-        if (vma.type != VMAType::Free) {
-            const auto backing_memory = memory.GetFCRAMOffset(vma.backing_memory + (interval_target - vma.base));
-            backing_blocks +=
-                MemoryRegionInfo::Interval(backing_memory, backing_memory + interval_size);
-        }
+        VAddr interval_end = std::min(address + size, vma->second.base + vma->second.size);
+        u32 interval_size = interval_end - interval_target;
+        auto backing_memory = memory.GetFCRAMOffset(vma->second.backing_memory +
+                                                    (interval_target - vma->second.base));
+        backing_blocks +=
+            MemoryRegionInfo::Interval(backing_memory, backing_memory + interval_size);

        interval_target += interval_size;
    }
    return backing_blocks;
 }
-
-ResultVal<VMManager::VMAHandle> VMManager::AliasMemory(VAddr target, VAddr source, u32 size,
-                                                       MemoryState alias_state,
-                                                       MemoryState aliased_state) {
-    VMAIter source_iter = StripIterConstness(FindVMA(source));
-    ASSERT(source_iter != vma_map.end());
-
-    auto& source_vma = source_iter->second;
-    ASSERT(source_vma.type == VMAType::BackingMemory);
-    ASSERT(source_vma.backing_memory != nullptr);
-    ASSERT(source + size <= source_vma.base + source_vma.size);
-
-    CASCADE_RESULT(VMAIter vma_handle, CarveVMA(target, size));
-    VirtualMemoryArea& final_vma = vma_handle->second;
-    ASSERT(final_vma.size == size);
-
-    final_vma.type = VMAType::MemoryAlias;
-    final_vma.permissions = VMAPermission::None;
-    final_vma.meminfo_state = alias_state;
-    final_vma.backing_memory = source_vma.backing_memory + (source - source_vma.base);
-    UpdatePageTableForVMA(final_vma);
-
-    // Update the source VMA state with the new value
-    CASCADE_RESULT(auto new_vma, CarveVMARange(source, size));
-    new_vma->second.meminfo_state = aliased_state;
-    UpdatePageTableForVMA(new_vma->second);
-
-    return MergeAdjacent(vma_handle);
-}
-
-ResultCode VMManager::AliasMemory(VAddr target, VAddr source, u32 size, MemoryState alias_state,
-                                  VMAPermission alias_permissions,
-                                  const VMManager& source_address_space) {
-
-    u32 cursor = 0;
-
-    while (true) {
-        VMAHandle source_iter = source_address_space.FindVMA(source + cursor);
-        ASSERT(source_iter != vma_map.end());
-
-        auto& source_vma = source_iter->second;
-        // Skip non-allocated blocks
-        if (source_vma.type == VMAType::Free) {
-            cursor += source_vma.size;
-            continue;
-        }
-
-        ASSERT(source_vma.type == VMAType::BackingMemory ||
-               source_vma.type == VMAType::MemoryAlias);
-        ASSERT(source_vma.backing_memory != nullptr);
-        ASSERT(source_vma.base == source + cursor);
-
-        if (source_vma.size + cursor > size) {
-            break;
-        }
-
-        CASCADE_RESULT(VMAIter vma_handle, CarveVMA(target + cursor, source_vma.size));
-        VirtualMemoryArea& final_vma = vma_handle->second;
-        ASSERT(final_vma.size == source_vma.size);
-
-        final_vma.type = VMAType::MemoryAlias;
-        final_vma.meminfo_state = alias_state;
-        final_vma.permissions = alias_permissions;
-        final_vma.backing_memory = source_vma.backing_memory;
-        UpdatePageTableForVMA(final_vma);
-
-        MergeAdjacent(vma_handle);
-
-        cursor += source_vma.size;
-    }
-
-    return RESULT_SUCCESS;
-}
-
 } // namespace Kernel
--- a/src/core/hle/kernel/vm_manager.h
+++ b/src/core/hle/kernel/vm_manager.h
@@ -27,8 +27,6 @@ enum class VMAType : u8 {
    BackingMemory,
    /// VMA is mapped to MMIO registers at a fixed PAddr.
    MMIO,
-    /// VMA is an alias of another VMA.
-    MemoryAlias,
 };

 /// Permissions for mapped memory blocks
@@ -180,12 +178,6 @@ public:
    ResultVal<VMAHandle> MapMMIO(VAddr target, PAddr paddr, u32 size, MemoryState state,
                                 Memory::MMIORegionPointer mmio_handler);

-    ResultVal<VMAHandle> AliasMemory(VAddr target, VAddr source, u32 size, MemoryState alias_state,
-                                     MemoryState aliased_state);
-
-    ResultCode AliasMemory(VAddr target, VAddr source, u32 size, MemoryState alias_state,
-                           VMAPermission alias_permissions, const VMManager& source_address_space);
-
    /**
     * Updates the memory state and permissions of the specified range. The range's original memory
     * state and permissions must match the `expected` parameters.
--- a/src/core/hle/service/fs/file.cpp
+++ b/src/core/hle/service/fs/file.cpp
@@ -57,7 +57,6 @@ void File::Read(Kernel::HLERequestContext& ctx) {
    IPC::RequestParser rp(ctx);
    u64 offset = rp.Pop<u64>();
    u32 length = rp.Pop<u32>();
-    auto& buffer = rp.PopMappedBuffer();
    LOG_TRACE(Service_FS, "Read {}: offset=0x{:x} length=0x{:08X}", GetName(), offset, length);

    const FileSessionSlot* file = GetSessionData(ctx.Session());
@@ -76,22 +75,94 @@ void File::Read(Kernel::HLERequestContext& ctx) {
                  offset, length, backend->GetSize());
    }

-    IPC::RequestBuilder rb = rp.MakeBuilder(2, 2);
+    // Conventional reading if the backend does not support cache.
+    if (!backend->AllowsCachedReads()) {
+        auto& buffer = rp.PopMappedBuffer();
+        IPC::RequestBuilder rb = rp.MakeBuilder(2, 2);
+        std::unique_ptr<u8*> data = std::make_unique<u8*>(static_cast<u8*>(operator new(length)));
+        const auto read = backend->Read(offset, length, *data);
+        if (read.Failed()) {
+            rb.Push(read.Code());
+            rb.Push<u32>(0);
+        } else {
+            buffer.Write(*data, 0, *read);
+            rb.Push(RESULT_SUCCESS);
+            rb.Push<u32>(static_cast<u32>(*read));
+        }
+        rb.PushMappedBuffer(buffer);

-    std::vector<u8> data(length);
-    ResultVal<std::size_t> read = backend->Read(offset, data.size(), data.data());
-    if (read.Failed()) {
-        rb.Push(read.Code());
-        rb.Push<u32>(0);
-    } else {
-        buffer.Write(data.data(), 0, *read);
-        rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(static_cast<u32>(*read));
+        std::chrono::nanoseconds read_timeout_ns{backend->GetReadDelayNs(length)};
+        ctx.SleepClientThread("file::read", read_timeout_ns, nullptr);
+        return;
    }
-    rb.PushMappedBuffer(buffer);

-    std::chrono::nanoseconds read_timeout_ns{backend->GetReadDelayNs(length)};
-    ctx.SleepClientThread("file::read", read_timeout_ns, nullptr);
+    struct AsyncData {
+        // Input
+        u32 length;
+        u64 offset;
+        std::chrono::steady_clock::time_point pre_timer;
+        bool cache_ready;
+
+        // Output
+        ResultCode ret{0};
+        Kernel::MappedBuffer* buffer;
+        std::unique_ptr<u8*> data;
+        size_t read_size;
+    };
+
+    auto async_data = std::make_shared<AsyncData>();
+    async_data->buffer = &rp.PopMappedBuffer();
+    async_data->length = length;
+    async_data->offset = offset;
+    async_data->cache_ready = backend->CacheReady(offset, length);
+    if (!async_data->cache_ready) {
+        async_data->pre_timer = std::chrono::steady_clock::now();
+    }
+
+    // LOG_DEBUG(Service_FS, "cache={}, offset={}, length={}", cache_ready, offset, length);
+    ctx.RunAsync(
+        [this, async_data](Kernel::HLERequestContext& ctx) {
+            async_data->data =
+                std::make_unique<u8*>(static_cast<u8*>(operator new(async_data->length)));
+            const auto read =
+                backend->Read(async_data->offset, async_data->length, *async_data->data);
+            if (read.Failed()) {
+                async_data->ret = read.Code();
+                async_data->read_size = 0;
+            } else {
+                async_data->ret = RESULT_SUCCESS;
+                async_data->read_size = *read;
+            }
+
+            const auto read_delay = static_cast<s64>(backend->GetReadDelayNs(async_data->length));
+            if (!async_data->cache_ready) {
+                const auto time_took = std::chrono::duration_cast<std::chrono::nanoseconds>(
+                                           std::chrono::steady_clock::now() - async_data->pre_timer)
+                                           .count();
+                /*
+                if (time_took > read_delay) {
+                    LOG_DEBUG(Service_FS, "Took longer! length={}, time_took={}, read_delay={}",
+                              async_data->length, time_took, read_delay);
+                }
+                */
+                return static_cast<s64>((read_delay > time_took) ? (read_delay - time_took) : 0);
+            } else {
+                return static_cast<s64>(read_delay);
+            }
+        },
+        [async_data](Kernel::HLERequestContext& ctx) {
+            IPC::RequestBuilder rb(ctx, 0x0802, 2, 2);
+            if (async_data->ret.IsError()) {
+                rb.Push(async_data->ret);
+                rb.Push<u32>(0);
+            } else {
+                async_data->buffer->Write(*async_data->data, 0, async_data->read_size);
+                rb.Push(RESULT_SUCCESS);
+                rb.Push<u32>(static_cast<u32>(async_data->read_size));
+            }
+            rb.PushMappedBuffer(*async_data->buffer);
+        },
+        !async_data->cache_ready);
 }

 void File::Write(Kernel::HLERequestContext& ctx) {
--- a/src/core/hle/service/fs/fs_user.cpp
+++ b/src/core/hle/service/fs/fs_user.cpp
@@ -1,8 +1,7 @@
 // Copyright 2014 Citra Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
-#pragma optimize("", off)
-#include <cryptopp/sha.h>
+
 #include "common/archives.h"
 #include "common/assert.h"
 #include "common/common_types.h"
@@ -671,6 +670,26 @@ void FS_USER::GetFormatInfo(Kernel::HLERequestContext& ctx) {
    rb.Push<bool>(format_info->duplicate_data != 0);
 }

+void FS_USER::GetProductInfo(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp(ctx);
+
+    u32 process_id = rp.Pop<u32>();
+
+    LOG_DEBUG(Service_FS, "called, process_id={}", process_id);
+
+    IPC::RequestBuilder rb = rp.MakeBuilder(6, 0);
+
+    const auto product_info = GetProductInfo(process_id);
+    if (!product_info.has_value()) {
+        rb.Push(ResultCode(FileSys::ErrCodes::ArchiveNotMounted, ErrorModule::FS,
+                           ErrorSummary::NotFound, ErrorLevel::Status));
+        return;
+    }
+
+    rb.Push(RESULT_SUCCESS);
+    rb.PushRaw<ProductInfo>(product_info.value());
+}
+
 void FS_USER::GetProgramLaunchInfo(Kernel::HLERequestContext& ctx) {
    IPC::RequestParser rp(ctx);
    const auto process_id = rp.Pop<u32>();
@@ -688,8 +707,20 @@ void FS_USER::GetProgramLaunchInfo(Kernel::HLERequestContext& ctx) {
        return;
    }

+    ProgramInfo program_info = program_info_result.Unwrap();
+
+    // Always report the launched program mediatype is SD if the friends module is requesting this
+    // information and the media type is game card. Otherwise, friends will append a "romid" field
+    // to the NASC request with a cartridge unique identifier. Using a dump of a game card and the
+    // game card itself at the same time online is known to have caused issues in the past.
+    auto process = ctx.ClientThread()->owner_process.lock();
+    if (process && process->codeset->name == "friends" &&
+        program_info.media_type == MediaType::GameCard) {
+        program_info.media_type = MediaType::SDMC;
+    }
+
    rb.Push(RESULT_SUCCESS);
-    rb.PushRaw(program_info_result.Unwrap());
+    rb.PushRaw<ProgramInfo>(program_info);
 }

 void FS_USER::ObsoletedCreateExtSaveData(Kernel::HLERequestContext& ctx) {
@@ -776,12 +807,12 @@ void FS_USER::AddSeed(Kernel::HLERequestContext& ctx) {
    rb.Push(RESULT_SUCCESS);
 }

-void FS_USER::SetSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
+void FS_USER::ObsoletedSetSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
    IPC::RequestParser rp(ctx);
-    u64 value = rp.Pop<u64>();
-    u32 secure_value_slot = rp.Pop<u32>();
-    u32 unique_id = rp.Pop<u32>();
-    u8 title_variation = rp.Pop<u8>();
+    const u64 value = rp.Pop<u64>();
+    const u32 secure_value_slot = rp.Pop<u32>();
+    const u32 unique_id = rp.Pop<u32>();
+    const u8 title_variation = rp.Pop<u8>();

    // TODO: Generate and Save the Secure Value

@@ -795,12 +826,11 @@ void FS_USER::SetSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
    rb.Push(RESULT_SUCCESS);
 }

-void FS_USER::GetSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
+void FS_USER::ObsoletedGetSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
    IPC::RequestParser rp(ctx);
-
-    u32 secure_value_slot = rp.Pop<u32>();
-    u32 unique_id = rp.Pop<u32>();
-    u8 title_variation = rp.Pop<u8>();
+    const u32 secure_value_slot = rp.Pop<u32>();
+    const u32 unique_id = rp.Pop<u32>();
+    const u8 title_variation = rp.Pop<u8>();

    LOG_WARNING(
        Service_FS,
@@ -817,7 +847,77 @@ void FS_USER::GetSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
    rb.Push<u64>(0);      // the secure value
 }

-void FS_USER::Register(u32 process_id, u64 program_id, const std::string& filepath) {
+void FS_USER::SetThisSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp(ctx);
+    const u32 secure_value_slot = rp.Pop<u32>();
+    const u64 value = rp.Pop<u64>();
+
+    // TODO: Generate and Save the Secure Value
+
+    LOG_WARNING(Service_FS, "(STUBBED) called, value=0x{:016x} secure_value_slot=0x{:08X}", value,
+                secure_value_slot);
+
+    IPC::RequestBuilder rb = rp.MakeBuilder(1, 0);
+
+    rb.Push(RESULT_SUCCESS);
+}
+
+void FS_USER::GetThisSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp(ctx);
+    const u32 secure_value_slot = rp.Pop<u32>();
+
+    LOG_WARNING(Service_FS, "(STUBBED) called secure_value_slot=0x{:08X}", secure_value_slot);
+
+    IPC::RequestBuilder rb = rp.MakeBuilder(5, 0);
+
+    rb.Push(RESULT_SUCCESS);
+
+    // TODO: Implement Secure Value Lookup & Generation
+
+    rb.Push<bool>(false); // indicates that the secure value doesn't exist
+    rb.Push<bool>(false); // looks like a boolean value, purpose unknown
+    rb.Push<u64>(0);      // the secure value
+}
+
+void FS_USER::SetSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp(ctx);
+    const auto archive_handle = rp.PopRaw<ArchiveHandle>();
+    const u32 secure_value_slot = rp.Pop<u32>();
+    const u64 value = rp.Pop<u64>();
+    const bool flush = rp.Pop<bool>();
+
+    // TODO: Generate and Save the Secure Value
+
+    LOG_WARNING(Service_FS,
+                "(STUBBED) called, value=0x{:016x} secure_value_slot=0x{:04X} "
+                "archive_handle=0x{:08X} flush={}",
+                value, secure_value_slot, archive_handle, flush);
+
+    IPC::RequestBuilder rb = rp.MakeBuilder(1, 0);
+
+    rb.Push(RESULT_SUCCESS);
+}
+
+void FS_USER::GetSaveDataSecureValue(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp(ctx);
+    const auto archive_handle = rp.PopRaw<ArchiveHandle>();
+    const u32 secure_value_slot = rp.Pop<u32>();
+
+    LOG_WARNING(Service_FS, "(STUBBED) called secure_value_slot=0x{:08X} archive_handle=0x{:08X}",
+                secure_value_slot, archive_handle);
+
+    IPC::RequestBuilder rb = rp.MakeBuilder(5, 0);
+
+    rb.Push(RESULT_SUCCESS);
+
+    // TODO: Implement Secure Value Lookup & Generation
+
+    rb.Push<bool>(false); // indicates that the secure value doesn't exist
+    rb.Push<bool>(false); // looks like a boolean value, purpose unknown
+    rb.Push<u64>(0);      // the secure value
+}
+
+void FS_USER::RegisterProgramInfo(u32 process_id, u64 program_id, const std::string& filepath) {
    const MediaType media_type = GetMediaTypeFromPath(filepath);
    program_info_map.insert_or_assign(process_id, ProgramInfo{program_id, media_type});
    if (media_type == MediaType::GameCard) {
@@ -829,6 +929,19 @@ std::string FS_USER::GetCurrentGamecardPath() const {
    return current_gamecard_path;
 }

+void FS_USER::RegisterProductInfo(u32 process_id, const ProductInfo& product_info) {
+    product_info_map.insert_or_assign(process_id, product_info);
+}
+
+std::optional<FS_USER::ProductInfo> FS_USER::GetProductInfo(u32 process_id) {
+    auto it = product_info_map.find(process_id);
+    if (it != product_info_map.end()) {
+        return it->second;
+    } else {
+        return {};
+    }
+}
+
 ResultVal<u16> FS_USER::GetSpecialContentIndexFromGameCard(u64 title_id, SpecialContentType type) {
    // TODO(B3N30) check if on real 3DS NCSD is checked if partition exists

@@ -879,24 +992,6 @@ ResultVal<u16> FS_USER::GetSpecialContentIndexFromTMD(MediaType media_type, u64
    return ResultCode(-1);
 }

-void FS_USER::UpdateSha256Context(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp(ctx);
-    rp.Skip(8, false);
-    [[maybe_unused]] const u32 size = rp.Pop<u32>();
-    rp.Skip(4, false);
-    auto input = rp.PopMappedBuffer();
-
-    std::array<u8, CryptoPP::SHA256::DIGESTSIZE> hash;
-    std::vector<u8> data(size);
-    input.Read(data.data(), 0, data.size());
-
-    CryptoPP::SHA256().CalculateDigest(hash.data(), data.data(), size);
-
-    IPC::RequestBuilder rb = rp.MakeBuilder(9, 0);
-    rb.Push(RESULT_SUCCESS);
-    rb.PushRaw(hash);
-}
-
 FS_USER::FS_USER(Core::System& system)
    : ServiceFramework("fs:USER", 30), system(system), archives(system.ArchiveManager()) {
    static const FunctionInfo functions[] = {
@@ -948,7 +1043,7 @@ FS_USER::FS_USER(Core::System& system)
        {0x082B, nullptr, "CardNorDirectRead_4xIO"},
        {0x082C, nullptr, "CardNorDirectCpuWriteWithoutVerify"},
        {0x082D, nullptr, "CardNorDirectSectorEraseWithoutVerify"},
-        {0x082E, nullptr, "GetProductInfo"},
+        {0x082E, &FS_USER::GetProductInfo, "GetProductInfo"},
        {0x082F, &FS_USER::GetProgramLaunchInfo, "GetProgramLaunchInfo"},
        {0x0830, &FS_USER::ObsoletedCreateExtSaveData, "Obsoleted_3_0_CreateExtSaveData"},
        {0x0831, nullptr, "CreateSharedExtSaveData"},
@@ -980,7 +1075,7 @@ FS_USER::FS_USER(Core::System& system)
        {0x084B, nullptr, "ImportIntegrityVerificationSeed"},
        {0x084C, &FS_USER::FormatSaveData, "FormatSaveData"},
        {0x084D, nullptr, "GetLegacySubBannerData"},
-        {0x084E, &FS_USER::UpdateSha256Context, "UpdateSha256Context"},
+        {0x084E, nullptr, "UpdateSha256Context"},
        {0x084F, nullptr, "ReadSpecialFile"},
        {0x0850, nullptr, "GetSpecialFileSize"},
        {0x0851, &FS_USER::CreateExtSaveData, "CreateExtSaveData"},
@@ -1003,12 +1098,16 @@ FS_USER::FS_USER(Core::System& system)
        {0x0862, &FS_USER::SetPriority, "SetPriority"},
        {0x0863, &FS_USER::GetPriority, "GetPriority"},
        {0x0864, nullptr, "GetNandInfo"},
-        {0x0865, &FS_USER::SetSaveDataSecureValue, "SetSaveDataSecureValue"},
-        {0x0866, &FS_USER::GetSaveDataSecureValue, "GetSaveDataSecureValue"},
+        {0x0865, &FS_USER::ObsoletedSetSaveDataSecureValue, "SetSaveDataSecureValue"},
+        {0x0866, &FS_USER::ObsoletedGetSaveDataSecureValue, "GetSaveDataSecureValue"},
        {0x0867, nullptr, "ControlSecureSave"},
        {0x0868, nullptr, "GetMediaType"},
        {0x0869, nullptr, "GetNandEraseCount"},
        {0x086A, nullptr, "ReadNandReport"},
+        {0x086E, &FS_USER::SetThisSaveDataSecureValue, "SetThisSaveDataSecureValue" },
+        {0x086F, &FS_USER::GetThisSaveDataSecureValue, "GetThisSaveDataSecureValue" },
+        {0x0875, &FS_USER::SetSaveDataSecureValue, "SetSaveDataSecureValue" },
+        {0x0876, &FS_USER::GetSaveDataSecureValue, "GetSaveDataSecureValue" },
        {0x087A, &FS_USER::AddSeed, "AddSeed"},
        {0x087D, &FS_USER::GetNumSeeds, "GetNumSeeds"},
        {0x0886, nullptr, "CheckUpdatedDat"},
--- a/src/core/hle/service/fs/fs_user.h
+++ b/src/core/hle/service/fs/fs_user.h
@@ -4,6 +4,7 @@

 #pragma once

+#include <optional>
 #include <unordered_map>
 #include <boost/serialization/base_object.hpp>
 #include "common/common_types.h"
@@ -48,12 +49,23 @@ class FS_USER final : public ServiceFramework<FS_USER, ClientSlot> {
 public:
    explicit FS_USER(Core::System& system);

-    // On real HW this is part of FS:Reg. But since that module is only used by loader and pm, which
-    // we HLEed, we can just directly use it here
-    void Register(u32 process_id, u64 program_id, const std::string& filepath);
+    // On real HW this is part of FSReg (FSReg:Register). But since that module is only used by
+    // loader and pm, which we HLEed, we can just directly use it here
+    void RegisterProgramInfo(u32 process_id, u64 program_id, const std::string& filepath);

    std::string GetCurrentGamecardPath() const;

+    struct ProductInfo {
+        std::array<u8, 0x10> product_code;
+        u16_le maker_code;
+        u16_le remaster_version;
+    };
+    static_assert(sizeof(ProductInfo) == 0x14);
+
+    void RegisterProductInfo(u32 process_id, const ProductInfo& product_info);
+
+    std::optional<ProductInfo> GetProductInfo(u32 process_id);
+
    /// Gets the registered program info of a process.
    ResultVal<ProgramInfo> GetProgramLaunchInfo(u32 process_id) const {
        auto info = program_info_map.find(process_id);
@@ -107,7 +119,7 @@ private:
     */
    void OpenFileDirectly(Kernel::HLERequestContext& ctx);

-    /**
+    /*
     * FS_User::DeleteFile service function
     *  Inputs:
     *      1 : Transaction
@@ -121,7 +133,7 @@ private:
     */
    void DeleteFile(Kernel::HLERequestContext& ctx);

-    /**
+    /*
     * FS_User::RenameFile service function
     *  Inputs:
     *      1 : Transaction
@@ -140,7 +152,7 @@ private:
     */
    void RenameFile(Kernel::HLERequestContext& ctx);

-    /**
+    /*
     * FS_User::DeleteDirectory service function
     *  Inputs:
     *      1 : Transaction
@@ -154,7 +166,7 @@ private:
     */
    void DeleteDirectory(Kernel::HLERequestContext& ctx);

-    /**
+    /*
     * FS_User::DeleteDirectoryRecursively service function
     *  Inputs:
     *      0 : Command header 0x08070142
@@ -169,7 +181,7 @@ private:
     */
    void DeleteDirectoryRecursively(Kernel::HLERequestContext& ctx);

-    /**
+    /*
     * FS_User::CreateFile service function
     *  Inputs:
     *      0 : Command header 0x08080202
@@ -186,7 +198,7 @@ private:
     */
    void CreateFile(Kernel::HLERequestContext& ctx);

-    /**
+    /*
     * FS_User::CreateDirectory service function
     *  Inputs:
     *      1 : Transaction
@@ -201,7 +213,7 @@ private:
     */
    void CreateDirectory(Kernel::HLERequestContext& ctx);

-    /**
+    /*
     * FS_User::RenameDirectory service function
     *  Inputs:
     *      1 : Transaction
@@ -510,6 +522,17 @@ private:
     */
    void GetFormatInfo(Kernel::HLERequestContext& ctx);

+    /**
+     * FS_User::GetProductInfo service function.
+     *  Inputs:
+     *      0 : 0x082E0040
+     *      1 : Process ID
+     *  Outputs:
+     *      1 : Result of function, 0 on success, otherwise error code
+     *      2-6 : Product info
+     */
+    void GetProductInfo(Kernel::HLERequestContext& ctx);
+
    /**
     * FS_User::GetProgramLaunchInfo service function.
     *  Inputs:
@@ -601,7 +624,7 @@ private:
     *      0 : 0x08650140
     *      1 : Result of function, 0 on success, otherwise error code
     */
-    void SetSaveDataSecureValue(Kernel::HLERequestContext& ctx);
+    void ObsoletedSetSaveDataSecureValue(Kernel::HLERequestContext& ctx);

    /**
     * FS_User::GetSaveDataSecureValue service function.
@@ -616,24 +639,58 @@ private:
     *      2 : If Secure Value doesn't exist, 0, if it exists, 1
     *      3-4 : Secure Value
     */
-    void GetSaveDataSecureValue(Kernel::HLERequestContext& ctx);
+    void ObsoletedGetSaveDataSecureValue(Kernel::HLERequestContext& ctx);

    /**
-     * FS_User::UpdateSha256Context service function.
+     * FS_User::SetThisSaveDataSecureValue service function.
     *  Inputs:
-     *      0 : 0x084E0342
-     *      1-8 : Input Hash
-     *      9 : Input Data Size
-     *      10-12 : Flag, must be zero
-     *      13 : Flag, must be non-zero
-     *      14 : (Size << 4) | 0xA
-     *      15 : Input Data Pointer
+     *      1 : Secure Value Slot
+     *      2-3 : Secure Value
     *  Outputs:
-     *      0 : 0x084E0342
     *      1 : Result of function, 0 on success, otherwise error code
-     *      2-9 : Output SHA-256 Hash
     */
-    void UpdateSha256Context(Kernel::HLERequestContext& ctx);
+    void SetThisSaveDataSecureValue(Kernel::HLERequestContext& ctx);
+
+    /**
+     * FS_User::GetSaveDataSecureValue service function.
+     *  Inputs:
+     *      1 : Secure Value Slot
+     *  Outputs:
+     *      1 : Result of function, 0 on success, otherwise error code
+     *      2 : If Secure Value doesn't exist, 0, if it exists, 1
+     *      3 : Unknown
+     *      4-5 : Secure Value
+     */
+    void GetThisSaveDataSecureValue(Kernel::HLERequestContext& ctx);
+
+    /**
+     * FS_User::SetSaveDataSecureValue service function.
+     *  Inputs:
+     *      0 : 0x08750180
+     *      1-2 : Archive
+     *      3 : Secure Value Slot
+     *      4 : value
+     *      5 : flush
+     *  Outputs:
+     *      0 : header
+     *      1 : Result of function, 0 on success, otherwise error code
+     */
+    void SetSaveDataSecureValue(Kernel::HLERequestContext& ctx);
+
+    /**
+     * FS_User::GetSaveDataSecureValue service function.
+     *  Inputs:
+     *      0 : 0x087600C0
+     *      1-2 : Archive
+     *      2 : Secure Value slot
+     *  Outputs:
+     *      0 : Header
+     *      1 : Result of function, 0 on success, otherwise error code
+     *      2 : If Secure Value doesn't exist, 0, if it exists, 1
+     *      3 : unknown
+     *      4-5 : Secure Value
+     */
+    void GetSaveDataSecureValue(Kernel::HLERequestContext& ctx);

    static ResultVal<u16> GetSpecialContentIndexFromGameCard(u64 title_id, SpecialContentType type);
    static ResultVal<u16> GetSpecialContentIndexFromTMD(MediaType media_type, u64 title_id,
@@ -642,6 +699,8 @@ private:
    std::unordered_map<u32, ProgramInfo> program_info_map;
    std::string current_gamecard_path;

+    std::unordered_map<u32, ProductInfo> product_info_map;
+
    u32 priority = -1; ///< For SetPriority and GetPriority service functions

    Core::System& system;
--- a/src/core/hle/service/ps/ps_ps.cpp
+++ b/src/core/hle/service/ps/ps_ps.cpp
@@ -1,10 +1,9 @@
 // Copyright 2018 Citra Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
-#pragma optimize("", off)
+
 #include <cryptopp/aes.h>
 #include <cryptopp/modes.h>
-#include <cryptopp/rsa.h>
 #include "common/archives.h"
 #include "common/logging/log.h"
 #include "core/core.h"
@@ -14,7 +13,6 @@
 #include "core/hw/aes/key.h"

 SERIALIZE_EXPORT_IMPL(Service::PS::PS_PS)
-SERVICE_CONSTRUCT_IMPL(Service::PS::PS_PS)

 namespace Service::PS {

@@ -148,50 +146,11 @@ void PS_PS::EncryptDecryptAes(Kernel::HLERequestContext& ctx) {
    rb.PushMappedBuffer(destination);
 }

-void PS_PS::VerifyRsaSha256(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp(ctx);
-    std::array<u8, 32> sha;
-    rp.PopRaw(sha);
-    rp.Pop<u32>();
-    const auto rsa_ctx_buf = rp.PopStaticBuffer();
-    auto sig_buf = rp.PopMappedBuffer();
-
-    struct RsaContext {
-        std::array<u8, 0x100> modulo;
-        std::array<u8, 0x100> exponent;
-        u32 rsa_bit_size;
-        bool is_long_exponent;
-    };
-    static_assert(sizeof(RsaContext) == 520);
-
-    RsaContext rsa_ctx;
-    std::memcpy(&rsa_ctx, rsa_ctx_buf.data(), rsa_ctx_buf.size());
-
-    std::vector<u8> signature_buffer(sig_buf.GetSize());
-    sig_buf.Read(signature_buffer.data(), 0, sig_buf.GetSize());
-
-    ASSERT(!rsa_ctx.is_long_exponent);
-
-    u32 exponent;
-    std::memcpy(&exponent, rsa_ctx.exponent.data(), sizeof(exponent));
-
-    using Verifier = CryptoPP::RSASS<CryptoPP::PKCS1v15, CryptoPP::SHA256>::Verifier;
-
-    const auto modulus = CryptoPP::Integer(rsa_ctx.modulo.data(), rsa_ctx.modulo.size());
-    const auto verifier = Verifier(modulus, CryptoPP::Integer(exponent));
-    const bool valid = verifier.VerifyMessage(sha.data(), sha.size(), signature_buffer.data(), signature_buffer.size());
-    ASSERT(!valid);
-
-    IPC::RequestBuilder rb = rp.MakeBuilder(1, 2);
-    rb.Push(RESULT_SUCCESS);
-    rb.PushMappedBuffer(sig_buf);
-}
-
-PS_PS::PS_PS(Core::System& system) : ServiceFramework("ps:ps", DefaultMaxSessions), memory{system.Memory()} {
+PS_PS::PS_PS() : ServiceFramework("ps:ps", DefaultMaxSessions) {
    static const FunctionInfo functions[] = {
        // clang-format off
        {0x0001, nullptr, "SignRsaSha256"},
-        {0x0002, &PS_PS::VerifyRsaSha256, "VerifyRsaSha256"},
+        {0x0002, nullptr, "VerifyRsaSha256"},
        {0x0004, &PS_PS::EncryptDecryptAes, "EncryptDecryptAes"},
        {0x0005, nullptr, "EncryptSignDecryptVerifyAesCcm"},
        {0x0006, nullptr, "GetRomId"},
@@ -214,7 +173,7 @@ PS_PS::PS_PS(Core::System& system) : ServiceFramework("ps:ps", DefaultMaxSession

 void InstallInterfaces(Core::System& system) {
    auto& service_manager = system.ServiceManager();
-    std::make_shared<PS_PS>(system)->InstallAsService(service_manager);
+    std::make_shared<PS_PS>()->InstallAsService(service_manager);
 }

 } // namespace Service::PS
--- a/src/core/hle/service/ps/ps_ps.h
+++ b/src/core/hle/service/ps/ps_ps.h
@@ -14,10 +14,11 @@ namespace Service::PS {

 class PS_PS final : public ServiceFramework<PS_PS> {
 public:
-    PS_PS(Core::System& system);
+    PS_PS();
    ~PS_PS() = default;

 private:
+    SERVICE_SERIALIZATION_SIMPLE

    /**
     * PS_PS::SignRsaSha256 service function
@@ -226,9 +227,6 @@ private:
     *      2 :
     */
    void InterfaceForPXI_0x04040044(Kernel::HLERequestContext& ctx);
-
-private:
-    Memory::MemorySystem& memory;
 };

 /// Initializes the PS_PS Service
@@ -237,4 +235,3 @@ void InstallInterfaces(Core::System& system);
 } // namespace Service::PS

 BOOST_CLASS_EXPORT_KEY(Service::PS::PS_PS)
-SERVICE_CONSTRUCT(Service::PS::PS_PS)
--- a/src/core/hle/service/service.cpp
+++ b/src/core/hle/service/service.cpp
@@ -58,8 +58,7 @@
 namespace Service {

 const std::array<ServiceModuleInfo, 41> service_module_map{
-    {{"CFG", 0x00040130'00001702, CFG::InstallInterfaces},
-     {"FS", 0x00040130'00001102, FS::InstallInterfaces},
+    {{"FS", 0x00040130'00001102, FS::InstallInterfaces},
     {"PM", 0x00040130'00001202, PM::InstallInterfaces},
     {"LDR", 0x00040130'00003702, LDR::InstallInterfaces},
     {"PXI", 0x00040130'00001402, PXI::InstallInterfaces},
@@ -75,6 +74,7 @@ const std::array<ServiceModuleInfo, 41> service_module_map{
          Y2R::InstallInterfaces(system);
      }},
     {"CECD", 0x00040130'00002602, CECD::InstallInterfaces},
+     {"CFG", 0x00040130'00001702, CFG::InstallInterfaces},
     {"DLP", 0x00040130'00002802, DLP::InstallInterfaces},
     {"DSP", 0x00040130'00001A02, DSP::InstallInterfaces},
     {"FRD", 0x00040130'00003202, FRD::InstallInterfaces},
--- a/src/core/loader/3dsx.cpp
+++ b/src/core/loader/3dsx.cpp
@@ -282,7 +282,7 @@ ResultStatus AppLoader_THREEDSX::Load(std::shared_ptr<Kernel::Process>& process)
    // On real HW this is done with FS:Reg, but we can be lazy
    auto fs_user =
        Core::System::GetInstance().ServiceManager().GetService<Service::FS::FS_USER>("fs:USER");
-    fs_user->Register(process->GetObjectId(), process->codeset->program_id, filepath);
+    fs_user->RegisterProgramInfo(process->GetObjectId(), process->codeset->program_id, filepath);

    process->Run(48, Kernel::DEFAULT_STACK_SIZE);

--- a/src/core/loader/ncch.cpp
+++ b/src/core/loader/ncch.cpp
@@ -174,7 +174,16 @@ ResultStatus AppLoader_NCCH::LoadExec(std::shared_ptr<Kernel::Process>& process)
        auto fs_user =
            Core::System::GetInstance().ServiceManager().GetService<Service::FS::FS_USER>(
                "fs:USER");
-        fs_user->Register(process->process_id, process->codeset->program_id, filepath);
+        fs_user->RegisterProgramInfo(process->process_id, process->codeset->program_id, filepath);
+
+        Service::FS::FS_USER::ProductInfo product_info{};
+        std::memcpy(product_info.product_code.data(), overlay_ncch->ncch_header.product_code,
+                    product_info.product_code.size());
+        std::memcpy(&product_info.remaster_version,
+                    overlay_ncch->exheader_header.codeset_info.flags.remaster_version,
+                    sizeof(product_info.remaster_version));
+        product_info.maker_code = overlay_ncch->ncch_header.maker_code;
+        fs_user->RegisterProductInfo(process->process_id, product_info);

        process->Run(priority, stack_size);
        return ResultStatus::Success;
--- a/src/core/memory.cpp
+++ b/src/core/memory.cpp
@@ -1,7 +1,7 @@
 // Copyright 2015 Citra Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
-#pragma optimize("", off)
+
 #include <array>
 #include <cstring>
 #include <boost/serialization/array.hpp>
@@ -156,10 +156,6 @@ public:
        return system.GetRunningCore().GetPC();
    }

-    u32 GetLR() const noexcept {
-        return system.GetRunningCore().GetReg(14);
-    }
-
    /**
     * This function should only be called for virtual addreses with attribute `PageType::Special`.
     */
@@ -473,8 +469,8 @@ T MemorySystem::Read(const VAddr vaddr) {
    PageType type = impl->current_page_table->attributes[vaddr >> CITRA_PAGE_BITS];
    switch (type) {
    case PageType::Unmapped:
-        LOG_ERROR(HW_Memory, "unmapped Read{} @ 0x{:08X} at PC 0x{:08X} with LR {:#08X}", sizeof(T) * 8, vaddr,
-                  impl->GetPC(), impl->GetLR());
+        LOG_ERROR(HW_Memory, "unmapped Read{} @ 0x{:08X} at PC 0x{:08X}", sizeof(T) * 8, vaddr,
+                  impl->GetPC());
        return 0;
    case PageType::Memory:
        ASSERT_MSG(false, "Mapped memory page without a pointer @ {:08X}", vaddr);
--- a/src/tests/video_core/shader/shader_jit_x64_compiler.cpp
+++ b/src/tests/video_core/shader/shader_jit_x64_compiler.cpp
@@ -11,6 +11,7 @@
 #include <span>
 #include <catch2/catch_approx.hpp>
 #include <catch2/catch_test_macros.hpp>
+#include <fmt/format.h>
 #include <nihstro/inline_assembly.h>
 #include "video_core/shader/shader_interpreter.h"
 #include "video_core/shader/shader_jit_x64_compiler.h"
@@ -28,6 +29,15 @@ static constexpr Common::Vec4f vec4_nan = Common::Vec4f::AssignToAll(NAN);
 static constexpr Common::Vec4f vec4_one = Common::Vec4f::AssignToAll(1.0f);
 static constexpr Common::Vec4f vec4_zero = Common::Vec4f::AssignToAll(0.0f);

+namespace Catch {
+template <>
+struct StringMaker<Common::Vec4f> {
+    static std::string convert(Common::Vec4f value) {
+        return fmt::format("({}, {}, {}, {})", value.r(), value.g(), value.b(), value.a());
+    }
+};
+} // namespace Catch
+
 static std::unique_ptr<Pica::Shader::ShaderSetup> CompileShaderSetup(
    std::initializer_list<nihstro::InlineAsm> code) {
    const auto shbin = nihstro::InlineAsm::CompileToRawBinary(code);
@@ -385,6 +395,56 @@ TEST_CASE("RSQ", "[video_core][shader][shader_jit]") {
    REQUIRE(shader.Run({0.0625f}).x == Catch::Approx(4.0f).margin(0.004f));
 }

+TEST_CASE("Address Register Offset", "[video_core][shader][shader_jit]") {
+    const auto sh_input = SourceRegister::MakeInput(0);
+    const auto sh_c40 = SourceRegister::MakeFloat(40);
+    const auto sh_output = DestRegister::MakeOutput(0);
+
+    auto shader = ShaderTest({
+        // mova a0.x, sh_input.x
+        {OpCode::Id::MOVA, DestRegister{}, "x", sh_input, "x", SourceRegister{}, "",
+         nihstro::InlineAsm::RelativeAddress::A1},
+        // mov sh_output.xyzw, c40[a0.x].xyzw
+        {OpCode::Id::MOV, sh_output, "xyzw", sh_c40, "xyzw", SourceRegister{}, "",
+         nihstro::InlineAsm::RelativeAddress::A1},
+        {OpCode::Id::END},
+    });
+
+    // Prepare shader uniforms
+    const bool inverted = true;
+    std::array<Common::Vec4f, 96> f_uniforms;
+    for (u32 i = 0; i < 0x80; i++) {
+        if (i >= 0x00 && i < 0x60) {
+            const u32 base = inverted ? (0x60 - i) : i;
+            const auto color = (base * 2.f) / 255.0f;
+            const auto color_f24 = Pica::f24::FromFloat32(color);
+            shader.shader_setup->uniforms.f[i] = {color_f24, color_f24, color_f24,
+                                                  Pica::f24::One()};
+            f_uniforms[i] = {color, color, color, 1.f};
+        } else if (i >= 0x60 && i < 0x64) {
+            const u8 color = static_cast<u8>((i - 0x60) * 0x10);
+            shader.shader_setup->uniforms.i[i - 0x60] = {color, color, color, 255};
+        } else if (i >= 0x70 && i < 0x80) {
+            shader.shader_setup->uniforms.b[i - 0x70] = i >= 0x78;
+        }
+    }
+
+    REQUIRE(shader.Run(0.f) == f_uniforms[40]);
+    REQUIRE(shader.Run(13.f) == f_uniforms[53]);
+    REQUIRE(shader.Run(50.f) == f_uniforms[90]);
+    REQUIRE(shader.Run(60.f) == vec4_one);
+    REQUIRE(shader.Run(74.f) == vec4_one);
+    REQUIRE(shader.Run(87.f) == vec4_one);
+    REQUIRE(shader.Run(88.f) == f_uniforms[0]);
+    REQUIRE(shader.Run(128.f) == f_uniforms[40]);
+    REQUIRE(shader.Run(-40.f) == f_uniforms[0]);
+    REQUIRE(shader.Run(-42.f) == vec4_one);
+    REQUIRE(shader.Run(-70.f) == vec4_one);
+    REQUIRE(shader.Run(-73.f) == f_uniforms[95]);
+    REQUIRE(shader.Run(-127.f) == f_uniforms[41]);
+    REQUIRE(shader.Run(-129.f) == f_uniforms[40]);
+}
+
 // TODO: Requires fix from https://github.com/neobrain/nihstro/issues/68
 // TEST_CASE("MAD", "[video_core][shader][shader_jit]") {
 //     const auto sh_input1 = SourceRegister::MakeInput(0);
--- a/src/video_core/command_processor.cpp
+++ b/src/video_core/command_processor.cpp
@@ -78,12 +78,11 @@ static void WriteUniformFloatReg(ShaderRegs& config, Shader::ShaderSetup& setup,
        (float_regs_counter >= 3 && !uniform_setup.IsFloat32())) {
        float_regs_counter = 0;

-        auto& uniform = setup.uniforms.f[uniform_setup.index];
-
-        if (uniform_setup.index >= 96) {
+        if (uniform_setup.index >= setup.uniforms.f.size()) {
            LOG_ERROR(HW_GPU, "Invalid {} float uniform index {}", GetShaderSetupTypeName(setup),
                      (int)uniform_setup.index);
        } else {
+            auto& uniform = setup.uniforms.f[uniform_setup.index];

            // NOTE: The destination component order indeed is "backwards"
            if (uniform_setup.IsFloat32()) {
--- a/src/video_core/host_shaders/CMakeLists.txt
+++ b/src/video_core/host_shaders/CMakeLists.txt
@@ -7,7 +7,6 @@ set(SHADER_FILES
    format_reinterpreter/rgba4_to_rgb5a1.frag
    format_reinterpreter/vulkan_d24s8_to_rgba8.comp
    texture_filtering/bicubic.frag
-    texture_filtering/nearest_neighbor.frag
    texture_filtering/refine.frag
    texture_filtering/scale_force.frag
    texture_filtering/xbrz_freescale.frag
--- a/src/video_core/host_shaders/texture_filtering/nearest_neighbor.frag
+++ b/src/video_core/host_shaders/texture_filtering/nearest_neighbor.frag
@@ -1,15 +0,0 @@
-// Copyright 2023 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-//? #version 430 core
-precision mediump float;
-
-layout(location = 0) in vec2 tex_coord;
-layout(location = 0) out vec4 frag_color;
-
-layout(binding = 2) uniform sampler2D input_texture;
-
-void main() {
-    frag_color = texture(input_texture, tex_coord);
-}
--- a/src/video_core/rasterizer_cache/rasterizer_cache.h
+++ b/src/video_core/rasterizer_cache/rasterizer_cache.h
@@ -361,10 +361,25 @@ typename T::Sampler& RasterizerCache<T>::GetSampler(SamplerId sampler_id) {
 template <class T>
 typename T::Sampler& RasterizerCache<T>::GetSampler(
    const Pica::TexturingRegs::TextureConfig& config) {
+    using TextureFilter = Pica::TexturingRegs::TextureConfig::TextureFilter;
+
+    const auto get_filter = [](TextureFilter filter) {
+        switch (Settings::values.texture_sampling.GetValue()) {
+        case Settings::TextureSampling::GameControlled:
+            return filter;
+        case Settings::TextureSampling::NearestNeighbor:
+            return TextureFilter::Nearest;
+        case Settings::TextureSampling::Linear:
+            return TextureFilter::Linear;
+        default:
+            return filter;
+        }
+    };
+
    const SamplerParams params = {
-        .mag_filter = config.mag_filter,
-        .min_filter = config.min_filter,
-        .mip_filter = config.mip_filter,
+        .mag_filter = get_filter(config.mag_filter),
+        .min_filter = get_filter(config.min_filter),
+        .mip_filter = get_filter(config.mip_filter),
        .wrap_s = config.wrap_s,
        .wrap_t = config.wrap_t,
        .border_color = config.border_color.raw,
--- a/src/video_core/renderer_base.cpp
+++ b/src/video_core/renderer_base.cpp
@@ -59,7 +59,7 @@ bool RendererBase::IsScreenshotPending() const {
    return settings.screenshot_requested;
 }

-void RendererBase::RequestScreenshot(void* data, std::function<void()> callback,
+void RendererBase::RequestScreenshot(void* data, std::function<void(bool)> callback,
                                     const Layout::FramebufferLayout& layout) {
    if (settings.screenshot_requested) {
        LOG_ERROR(Render, "A screenshot is already requested or in progress, ignoring the request");
--- a/src/video_core/renderer_base.h
+++ b/src/video_core/renderer_base.h
@@ -28,7 +28,7 @@ struct RendererSettings {
    // Screenshot
    std::atomic_bool screenshot_requested{false};
    void* screenshot_bits{};
-    std::function<void()> screenshot_complete_callback;
+    std::function<void(bool)> screenshot_complete_callback;
    Layout::FramebufferLayout screenshot_framebuffer_layout;
    // Renderer
    std::atomic_bool bg_color_update_requested{false};
@@ -103,7 +103,7 @@ public:
    [[nodiscard]] bool IsScreenshotPending() const;

    /// Request a screenshot of the next frame
-    void RequestScreenshot(void* data, std::function<void()> callback,
+    void RequestScreenshot(void* data, std::function<void(bool)> callback,
                           const Layout::FramebufferLayout& layout);

 protected:
--- a/src/video_core/renderer_opengl/gl_blit_helper.cpp
+++ b/src/video_core/renderer_opengl/gl_blit_helper.cpp
@@ -15,7 +15,6 @@
 #include "video_core/host_shaders/full_screen_triangle_vert.h"
 #include "video_core/host_shaders/texture_filtering/bicubic_frag.h"
 #include "video_core/host_shaders/texture_filtering/mmpx_frag.h"
-#include "video_core/host_shaders/texture_filtering/nearest_neighbor_frag.h"
 #include "video_core/host_shaders/texture_filtering/refine_frag.h"
 #include "video_core/host_shaders/texture_filtering/scale_force_frag.h"
 #include "video_core/host_shaders/texture_filtering/x_gradient_frag.h"
@@ -58,7 +57,6 @@ BlitHelper::BlitHelper(const Driver& driver_)
    : driver{driver_}, linear_sampler{CreateSampler(GL_LINEAR)},
      nearest_sampler{CreateSampler(GL_NEAREST)}, bicubic_program{CreateProgram(
                                                      HostShaders::BICUBIC_FRAG)},
-      nearest_program{CreateProgram(HostShaders::NEAREST_NEIGHBOR_FRAG)},
      scale_force_program{CreateProgram(HostShaders::SCALE_FORCE_FRAG)},
      xbrz_program{CreateProgram(HostShaders::XBRZ_FREESCALE_FRAG)},
      mmpx_program{CreateProgram(HostShaders::MMPX_FRAG)}, gradient_x_program{CreateProgram(
@@ -175,9 +173,6 @@ bool BlitHelper::Filter(Surface& surface, const VideoCore::TextureBlit& blit) {
    case TextureFilter::Bicubic:
        FilterBicubic(surface, blit);
        break;
-    case TextureFilter::NearestNeighbor:
-        FilterNearest(surface, blit);
-        break;
    case TextureFilter::ScaleForce:
        FilterScaleForce(surface, blit);
        break;
@@ -257,14 +252,6 @@ void BlitHelper::FilterBicubic(Surface& surface, const VideoCore::TextureBlit& b
    Draw(bicubic_program, surface.Handle(), draw_fbo.handle, blit.dst_level, blit.dst_rect);
 }

-void BlitHelper::FilterNearest(Surface& surface, const VideoCore::TextureBlit& blit) {
-    const OpenGLState prev_state = OpenGLState::GetCurState();
-    SCOPE_EXIT({ prev_state.Apply(); });
-    state.texture_units[2].texture_2d = surface.Handle(0);
-    SetParams(nearest_program, surface.RealExtent(false), blit.src_rect);
-    Draw(nearest_program, surface.Handle(), draw_fbo.handle, blit.dst_level, blit.dst_rect);
-}
-
 void BlitHelper::FilterScaleForce(Surface& surface, const VideoCore::TextureBlit& blit) {
    const OpenGLState prev_state = OpenGLState::GetCurState();
    SCOPE_EXIT({ prev_state.Apply(); });
--- a/src/video_core/renderer_opengl/gl_blit_helper.h
+++ b/src/video_core/renderer_opengl/gl_blit_helper.h
@@ -33,20 +33,13 @@ public:

 private:
    void FilterAnime4K(Surface& surface, const VideoCore::TextureBlit& blit);
-
    void FilterBicubic(Surface& surface, const VideoCore::TextureBlit& blit);
-
-    void FilterNearest(Surface& surface, const VideoCore::TextureBlit& blit);
-
    void FilterScaleForce(Surface& surface, const VideoCore::TextureBlit& blit);
-
    void FilterXbrz(Surface& surface, const VideoCore::TextureBlit& blit);
-
    void FilterMMPX(Surface& surface, const VideoCore::TextureBlit& blit);

    void SetParams(OGLProgram& program, const VideoCore::Extent& src_extent,
                   Common::Rectangle<u32> src_rect);
-
    void Draw(OGLProgram& program, GLuint dst_tex, GLuint dst_fbo, u32 dst_level,
              Common::Rectangle<u32> dst_rect);

@@ -59,7 +52,6 @@ private:
    OGLSampler nearest_sampler;

    OGLProgram bicubic_program;
-    OGLProgram nearest_program;
    OGLProgram scale_force_program;
    OGLProgram xbrz_program;
    OGLProgram mmpx_program;
--- a/src/video_core/renderer_opengl/renderer_opengl.cpp
+++ b/src/video_core/renderer_opengl/renderer_opengl.cpp
@@ -151,7 +151,7 @@ void RendererOpenGL::RenderScreenshot() {
        state.Apply();
        glDeleteRenderbuffers(1, &renderbuffer);

-        settings.screenshot_complete_callback();
+        settings.screenshot_complete_callback(true);
    }
 }

--- a/src/video_core/renderer_vulkan/renderer_vulkan.cpp
+++ b/src/video_core/renderer_vulkan/renderer_vulkan.cpp
@@ -6,6 +6,7 @@
 #include "common/logging/log.h"
 #include "common/microprofile.h"
 #include "common/settings.h"
+#include "common/texture.h"
 #include "core/core.h"
 #include "core/frontend/emu_window.h"
 #include "core/hw/gpu.h"
@@ -18,6 +19,7 @@
 #include "video_core/host_shaders/vulkan_present_frag_spv.h"
 #include "video_core/host_shaders/vulkan_present_interlaced_frag_spv.h"
 #include "video_core/host_shaders/vulkan_present_vert_spv.h"
+#include "vulkan/vulkan_format_traits.hpp"

 #include <vk_mem_alloc.h>

@@ -867,20 +869,9 @@ void RendererVulkan::RenderScreenshot() {
    const u32 width = layout.width;
    const u32 height = layout.height;

-    const vk::ImageCreateInfo staging_image_info = {
-        .imageType = vk::ImageType::e2D,
-        .format = vk::Format::eB8G8R8A8Unorm,
-        .extent{
-            .width = width,
-            .height = height,
-            .depth = 1,
-        },
-        .mipLevels = 1,
-        .arrayLayers = 1,
-        .samples = vk::SampleCountFlagBits::e1,
-        .tiling = vk::ImageTiling::eLinear,
-        .usage = vk::ImageUsageFlagBits::eTransferDst,
-        .initialLayout = vk::ImageLayout::eUndefined,
+    const vk::BufferCreateInfo staging_buffer_info = {
+        .size = width * height * 4,
+        .usage = vk::BufferUsageFlagBits::eTransferDst,
    };

    const VmaAllocationCreateInfo alloc_create_info = {
@@ -893,18 +884,18 @@ void RendererVulkan::RenderScreenshot() {
        .pUserData = nullptr,
    };

-    VkImage unsafe_image{};
+    VkBuffer unsafe_buffer{};
    VmaAllocation allocation{};
    VmaAllocationInfo alloc_info;
-    VkImageCreateInfo unsafe_image_info = static_cast<VkImageCreateInfo>(staging_image_info);
+    VkBufferCreateInfo unsafe_buffer_info = static_cast<VkBufferCreateInfo>(staging_buffer_info);

-    VkResult result = vmaCreateImage(instance.GetAllocator(), &unsafe_image_info,
-                                     &alloc_create_info, &unsafe_image, &allocation, &alloc_info);
+    VkResult result = vmaCreateBuffer(instance.GetAllocator(), &unsafe_buffer_info,
+                                      &alloc_create_info, &unsafe_buffer, &allocation, &alloc_info);
    if (result != VK_SUCCESS) [[unlikely]] {
        LOG_CRITICAL(Render_Vulkan, "Failed allocating texture with error {}", result);
        UNREACHABLE();
    }
-    vk::Image staging_image{unsafe_image};
+    vk::Buffer staging_buffer{unsafe_buffer};

    Frame frame{};
    main_window.RecreateFrame(&frame, width, height);
@@ -912,73 +903,37 @@ void RendererVulkan::RenderScreenshot() {
    DrawScreens(&frame, layout, false);

    scheduler.Record(
-        [width, height, source_image = frame.image, staging_image](vk::CommandBuffer cmdbuf) {
-            const std::array read_barriers = {
-                vk::ImageMemoryBarrier{
-                    .srcAccessMask = vk::AccessFlagBits::eMemoryWrite,
-                    .dstAccessMask = vk::AccessFlagBits::eTransferRead,
-                    .oldLayout = vk::ImageLayout::eTransferSrcOptimal,
-                    .newLayout = vk::ImageLayout::eTransferSrcOptimal,
-                    .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
-                    .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
-                    .image = source_image,
-                    .subresourceRange{
-                        .aspectMask = vk::ImageAspectFlagBits::eColor,
-                        .baseMipLevel = 0,
-                        .levelCount = VK_REMAINING_MIP_LEVELS,
-                        .baseArrayLayer = 0,
-                        .layerCount = VK_REMAINING_ARRAY_LAYERS,
-                    },
-                },
-                vk::ImageMemoryBarrier{
-                    .srcAccessMask = vk::AccessFlagBits::eNone,
-                    .dstAccessMask = vk::AccessFlagBits::eTransferWrite,
-                    .oldLayout = vk::ImageLayout::eUndefined,
-                    .newLayout = vk::ImageLayout::eTransferDstOptimal,
-                    .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
-                    .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
-                    .image = staging_image,
-                    .subresourceRange{
-                        .aspectMask = vk::ImageAspectFlagBits::eColor,
-                        .baseMipLevel = 0,
-                        .levelCount = VK_REMAINING_MIP_LEVELS,
-                        .baseArrayLayer = 0,
-                        .layerCount = VK_REMAINING_ARRAY_LAYERS,
-                    },
+        [width, height, source_image = frame.image, staging_buffer](vk::CommandBuffer cmdbuf) {
+            const vk::ImageMemoryBarrier read_barrier = {
+                .srcAccessMask = vk::AccessFlagBits::eMemoryWrite,
+                .dstAccessMask = vk::AccessFlagBits::eTransferRead,
+                .oldLayout = vk::ImageLayout::eTransferSrcOptimal,
+                .newLayout = vk::ImageLayout::eTransferSrcOptimal,
+                .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+                .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+                .image = source_image,
+                .subresourceRange{
+                    .aspectMask = vk::ImageAspectFlagBits::eColor,
+                    .baseMipLevel = 0,
+                    .levelCount = VK_REMAINING_MIP_LEVELS,
+                    .baseArrayLayer = 0,
+                    .layerCount = VK_REMAINING_ARRAY_LAYERS,
                },
            };
-            const std::array write_barriers = {
-                vk::ImageMemoryBarrier{
-                    .srcAccessMask = vk::AccessFlagBits::eTransferRead,
-                    .dstAccessMask = vk::AccessFlagBits::eMemoryWrite,
-                    .oldLayout = vk::ImageLayout::eTransferSrcOptimal,
-                    .newLayout = vk::ImageLayout::eTransferSrcOptimal,
-                    .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
-                    .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
-                    .image = source_image,
-                    .subresourceRange{
-                        .aspectMask = vk::ImageAspectFlagBits::eColor,
-                        .baseMipLevel = 0,
-                        .levelCount = VK_REMAINING_MIP_LEVELS,
-                        .baseArrayLayer = 0,
-                        .layerCount = VK_REMAINING_ARRAY_LAYERS,
-                    },
-                },
-                vk::ImageMemoryBarrier{
-                    .srcAccessMask = vk::AccessFlagBits::eTransferWrite,
-                    .dstAccessMask = vk::AccessFlagBits::eMemoryRead,
-                    .oldLayout = vk::ImageLayout::eTransferDstOptimal,
-                    .newLayout = vk::ImageLayout::eGeneral,
-                    .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
-                    .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
-                    .image = staging_image,
-                    .subresourceRange{
-                        .aspectMask = vk::ImageAspectFlagBits::eColor,
-                        .baseMipLevel = 0,
-                        .levelCount = VK_REMAINING_MIP_LEVELS,
-                        .baseArrayLayer = 0,
-                        .layerCount = VK_REMAINING_ARRAY_LAYERS,
-                    },
+            const vk::ImageMemoryBarrier write_barrier = {
+                .srcAccessMask = vk::AccessFlagBits::eTransferRead,
+                .dstAccessMask = vk::AccessFlagBits::eMemoryWrite,
+                .oldLayout = vk::ImageLayout::eTransferSrcOptimal,
+                .newLayout = vk::ImageLayout::eTransferSrcOptimal,
+                .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+                .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+                .image = source_image,
+                .subresourceRange{
+                    .aspectMask = vk::ImageAspectFlagBits::eColor,
+                    .baseMipLevel = 0,
+                    .levelCount = VK_REMAINING_MIP_LEVELS,
+                    .baseArrayLayer = 0,
+                    .layerCount = VK_REMAINING_ARRAY_LAYERS,
                },
            };
            static constexpr vk::MemoryBarrier memory_write_barrier = {
@@ -986,41 +941,29 @@ void RendererVulkan::RenderScreenshot() {
                .dstAccessMask = vk::AccessFlagBits::eMemoryRead | vk::AccessFlagBits::eMemoryWrite,
            };

-            const std::array src_offsets = {
-                vk::Offset3D{0, 0, 0},
-                vk::Offset3D{static_cast<s32>(width), static_cast<s32>(height), 1},
-            };
-
-            const std::array dst_offsets = {
-                vk::Offset3D{0, static_cast<s32>(height), 0},
-                vk::Offset3D{static_cast<s32>(width), 0, 1},
-            };
-
-            const vk::ImageBlit blit_area = {
-                .srcSubresource{
-                    .aspectMask = vk::ImageAspectFlagBits::eColor,
-                    .mipLevel = 0,
-                    .baseArrayLayer = 0,
-                    .layerCount = 1,
-                },
-                .srcOffsets = src_offsets,
-                .dstSubresource{
-                    .aspectMask = vk::ImageAspectFlagBits::eColor,
-                    .mipLevel = 0,
-                    .baseArrayLayer = 0,
-                    .layerCount = 1,
-                },
-                .dstOffsets = dst_offsets,
+            const vk::BufferImageCopy image_copy = {
+                .bufferOffset = 0,
+                .bufferRowLength = 0,
+                .bufferImageHeight = 0,
+                .imageSubresource =
+                    {
+                        .aspectMask = vk::ImageAspectFlagBits::eColor,
+                        .mipLevel = 0,
+                        .baseArrayLayer = 0,
+                        .layerCount = 1,
+                    },
+                .imageOffset = {0, 0, 0},
+                .imageExtent = {width, height, 1},
            };

            cmdbuf.pipelineBarrier(vk::PipelineStageFlagBits::eAllCommands,
                                   vk::PipelineStageFlagBits::eTransfer,
-                                   vk::DependencyFlagBits::eByRegion, {}, {}, read_barriers);
-            cmdbuf.blitImage(source_image, vk::ImageLayout::eTransferSrcOptimal, staging_image,
-                             vk::ImageLayout::eTransferDstOptimal, blit_area, vk::Filter::eNearest);
+                                   vk::DependencyFlagBits::eByRegion, {}, {}, read_barrier);
+            cmdbuf.copyImageToBuffer(source_image, vk::ImageLayout::eTransferSrcOptimal,
+                                     staging_buffer, image_copy);
            cmdbuf.pipelineBarrier(
                vk::PipelineStageFlagBits::eTransfer, vk::PipelineStageFlagBits::eAllCommands,
-                vk::DependencyFlagBits::eByRegion, memory_write_barrier, {}, write_barriers);
+                vk::DependencyFlagBits::eByRegion, memory_write_barrier, {}, write_barrier);
        });

    // Ensure the copy is fully completed before saving the screenshot
@@ -1028,27 +971,16 @@ void RendererVulkan::RenderScreenshot() {

    const vk::Device device = instance.GetDevice();

-    // Get layout of the image (including row pitch)
-    const vk::ImageSubresource subresource = {
-        .aspectMask = vk::ImageAspectFlagBits::eColor,
-        .mipLevel = 0,
-        .arrayLayer = 0,
-    };
-
-    const vk::SubresourceLayout subresource_layout =
-        device.getImageSubresourceLayout(staging_image, subresource);
-
    // Copy backing image data to the QImage screenshot buffer
-    const u8* data = reinterpret_cast<const u8*>(alloc_info.pMappedData);
-    std::memcpy(settings.screenshot_bits, data + subresource_layout.offset,
-                subresource_layout.size);
+    std::memcpy(settings.screenshot_bits, alloc_info.pMappedData, staging_buffer_info.size);

    // Destroy allocated resources
+    vmaDestroyBuffer(instance.GetAllocator(), unsafe_buffer, allocation);
    vmaDestroyImage(instance.GetAllocator(), frame.image, frame.allocation);
    device.destroyFramebuffer(frame.framebuffer);
    device.destroyImageView(frame.image_view);

-    settings.screenshot_complete_callback();
+    settings.screenshot_complete_callback(false);
 }

 } // namespace Vulkan
--- a/src/video_core/renderer_vulkan/vk_blit_helper.h
+++ b/src/video_core/renderer_vulkan/vk_blit_helper.h
@@ -35,10 +35,7 @@ public:
                       const VideoCore::BufferTextureCopy& copy);

 private:
-    /// Creates compute pipelines used for blit
    vk::Pipeline MakeComputePipeline(vk::ShaderModule shader, vk::PipelineLayout layout);
-
-    /// Creates graphics pipelines used for blit
    vk::Pipeline MakeDepthStencilBlitPipeline();

 private:
--- a/src/video_core/renderer_vulkan/vk_swapchain.cpp
+++ b/src/video_core/renderer_vulkan/vk_swapchain.cpp
@@ -150,8 +150,7 @@ void Swapchain::FindPresentFormat() {
        return;
    }

-    LOG_CRITICAL(Render_Vulkan, "Unable to find required swapchain format!");
-    UNREACHABLE();
+    UNREACHABLE_MSG("Unable to find required swapchain format!");
 }

 void Swapchain::SetPresentMode() {
--- a/src/video_core/renderer_vulkan/vk_texture_runtime.cpp
+++ b/src/video_core/renderer_vulkan/vk_texture_runtime.cpp
@@ -1562,7 +1562,7 @@ DebugScope::DebugScope(TextureRuntime& runtime, Common::Vec4f color, std::string
    if (!has_debug_tool) {
        return;
    }
-    scheduler.Record([color, label](vk::CommandBuffer cmdbuf) {
+    scheduler.Record([color, label = std::string(label)](vk::CommandBuffer cmdbuf) {
        const vk::DebugUtilsLabelEXT debug_label = {
            .pLabelName = label.data(),
            .color = std::array{color[0], color[1], color[2], color[3]},
--- a/src/video_core/shader/shader_jit_x64_compiler.cpp
+++ b/src/video_core/shader/shader_jit_x64_compiler.cpp
@@ -232,21 +232,45 @@ void JitShader::Compile_SwizzleSrc(Instruction instr, unsigned src_num, SourceRe
        address_register_index = instr.common.address_register_index;
    }

-    if (src_num == offset_src && address_register_index != 0) {
+    if (src_reg.GetRegisterType() == RegisterType::FloatUniform && src_num == offset_src &&
+        address_register_index != 0) {
+        Xbyak::Reg64 address_reg;
        switch (address_register_index) {
-        case 1: // address offset 1
-            movaps(dest, xword[src_ptr + ADDROFFS_REG_0 + src_offset_disp]);
+        case 1:
+            address_reg = ADDROFFS_REG_0;
            break;
-        case 2: // address offset 2
-            movaps(dest, xword[src_ptr + ADDROFFS_REG_1 + src_offset_disp]);
+        case 2:
+            address_reg = ADDROFFS_REG_1;
            break;
-        case 3: // address offset 3
-            movaps(dest, xword[src_ptr + LOOPCOUNT_REG.cvt64() + src_offset_disp]);
+        case 3:
+            address_reg = LOOPCOUNT_REG.cvt64();
            break;
        default:
            UNREACHABLE();
            break;
        }
+        // s32 offset = address_reg >= -128 && address_reg <= 127 ? address_reg : 0;
+        // u32 index = (src_reg.GetIndex() + offset) & 0x7f;
+
+        // First we add 128 to address_reg so the first comparison is turned to
+        // address_reg >= 0 && address_reg < 256 which can be performed with
+        // a single unsigned comparison (cmovb)
+        lea(eax, ptr[address_reg + 128]);
+        mov(ebx, src_reg.GetIndex());
+        mov(ecx, address_reg.cvt32());
+        add(ecx, ebx);
+        cmp(eax, 256);
+        cmovb(ebx, ecx);
+        and_(ebx, 0x7f);
+
+        // index > 95 ? vec4(1.0) : uniforms.f[index];
+        movaps(dest, ONE);
+        cmp(ebx, 95);
+        Label load_end;
+        jg(load_end);
+        shl(rbx, 4);
+        movaps(dest, xword[src_ptr + rbx]);
+        L(load_end);
    } else {
        // Load the source
        movaps(dest, xword[src_ptr + src_offset_disp]);
@@ -590,24 +614,14 @@ void JitShader::Compile_MOVA(Instruction instr) {
        // Move and sign-extend high 32 bits
        shr(rax, 32);
        movsxd(ADDROFFS_REG_1, eax);
-
-        // Multiply by 16 to be used as an offset later
-        shl(ADDROFFS_REG_0, 4);
-        shl(ADDROFFS_REG_1, 4);
    } else {
        if (swiz.DestComponentEnabled(0)) {
            // Move and sign-extend low 32 bits
            movsxd(ADDROFFS_REG_0, eax);
-
-            // Multiply by 16 to be used as an offset later
-            shl(ADDROFFS_REG_0, 4);
        } else if (swiz.DestComponentEnabled(1)) {
            // Move and sign-extend high 32 bits
            shr(rax, 32);
            movsxd(ADDROFFS_REG_1, eax);
-
-            // Multiply by 16 to be used as an offset later
-            shl(ADDROFFS_REG_1, 4);
        }
    }
 }
@@ -659,9 +673,6 @@ void JitShader::Compile_END(Instruction instr) {
    mov(byte[STATE + offsetof(UnitState, conditional_code[1])], COND1.cvt8());

    // Save address/loop registers
-    sar(ADDROFFS_REG_0, 4);
-    sar(ADDROFFS_REG_1, 4);
-    sar(LOOPCOUNT_REG, 4);
    mov(dword[STATE + offsetof(UnitState, address_registers[0])], ADDROFFS_REG_0.cvt32());
    mov(dword[STATE + offsetof(UnitState, address_registers[1])], ADDROFFS_REG_1.cvt32());
    mov(dword[STATE + offsetof(UnitState, address_registers[2])], LOOPCOUNT_REG);
@@ -813,11 +824,11 @@ void JitShader::Compile_LOOP(Instruction instr) {
    std::size_t offset = Uniforms::GetIntUniformOffset(instr.flow_control.int_uniform_id);
    mov(LOOPCOUNT, dword[UNIFORMS + offset]);
    mov(LOOPCOUNT_REG, LOOPCOUNT);
-    shr(LOOPCOUNT_REG, 4);
-    and_(LOOPCOUNT_REG, 0xFF0); // Y-component is the start
+    shr(LOOPCOUNT_REG, 8);
+    and_(LOOPCOUNT_REG, 0xFF); // Y-component is the start
    mov(LOOPINC, LOOPCOUNT);
-    shr(LOOPINC, 12);
-    and_(LOOPINC, 0xFF0);               // Z-component is the incrementer
+    shr(LOOPINC, 16);
+    and_(LOOPINC, 0xFF);                // Z-component is the incrementer
    movzx(LOOPCOUNT, LOOPCOUNT.cvt8()); // X-component is iteration count
    add(LOOPCOUNT, 1);                  // Iteration count is X-component + 1

@@ -993,9 +1004,6 @@ void JitShader::Compile(const std::array<u32, MAX_PROGRAM_CODE_LENGTH>* program_
    movsxd(ADDROFFS_REG_0, dword[STATE + offsetof(UnitState, address_registers[0])]);
    movsxd(ADDROFFS_REG_1, dword[STATE + offsetof(UnitState, address_registers[1])]);
    mov(LOOPCOUNT_REG, dword[STATE + offsetof(UnitState, address_registers[2])]);
-    shl(ADDROFFS_REG_0, 4);
-    shl(ADDROFFS_REG_1, 4);
-    shl(LOOPCOUNT_REG, 4);

    // Load conditional code
    mov(COND0, byte[STATE + offsetof(UnitState, conditional_code[0])]);
Author	SHA1	Message	Date
GPUCode	b94374009e	code: Add texture sampling option * This replaces the nearest neighbour filter that shouldn't have existed in the first place	2023-11-23 01:22:47 +02:00
Wunk	b10f3d96f5	command_processor: Fix out-of-bounds float-uniform access (#7111 ) Addresses: https://github.com/citra-emu/citra/issues/6696 https://github.com/citra-emu/citra/issues/6871	2023-11-03 03:35:52 -07:00
Steveice10	b5d744bcae	ci: Work around macOS GitHub runner pip install failures. (#7110 )	2023-11-03 03:35:32 -07:00
Castor215	89d5d4a2b6	externals: allow user to use system cubeb (#7107 )	2023-11-02 17:33:40 -07:00
PabloMK7	4284893044	Implement RomFS cache and async reads. (#7089 ) * Implement RomFS cache and async reads. * Suggestions and fix compilation. * Apply suggestions	2023-11-02 17:19:00 -07:00
Steveice10	79ea06b226	qt: Update to 6.6.0 (#7099 )	2023-11-01 17:58:02 -07:00
Castor215	8d811913a5	externals: allow user to use system cryptopp (#7105 )	2023-11-01 17:57:10 -07:00
Wunk	ac9d72a95c	vk_texture_runtime: Fix debug scope label lambda-capture (#7102 ) `DebugScope` was capturing a `string_view` in a lambda which is only valid during the scope of this ctor. When the lambda gets invoked at a later time, it will read undefined garbage. The lambda needs to make a deep copy of this `string_view` into a `string` so that it is valid by the time the scheduler invokes this lambda.	2023-11-01 21:30:54 +01:00
Castor215	d3ce43782d	externals: allow users to use system libenet (#7100 )	2023-10-31 14:01:50 -07:00
PabloMK7	597a2e8ead	Add missing FS:USER functions (#7051 )	2023-10-31 14:01:25 -07:00
TGP17	b231a22ea5	Switch compiler to clang on Linux (#7077 )	2023-10-25 11:00:52 -07:00
Steveice10	1110c01657	ci: Install Vulkan SPIRV-Tools on Windows to fix glslang install error. (#7092 )	2023-10-25 11:00:10 -07:00
Steveice10	45ef11654a	audio_core: Clear time stretcher after flushing to avoid sample bleed. (#7081 )	2023-10-24 17:22:10 -07:00
Dominik Kreutzer	259dbf17dc	citra-qt: ensure image interface is registered before starting game (#7090 )	2023-10-24 17:21:38 -07:00
Castor215	ec55807669	build: fix build failure when not using precompiled headers (#7087 ) Co-authored-by: vitor-k <vitor-kiguchi@hotmail.com>	2023-10-23 17:21:35 -03:00
GPUCode	36146459f8	renderer_vulkan: Fix screenshots under NVIDIA vulkan (#7082 )	2023-10-22 22:53:14 +03:00
Wunk	597297ffb4	tests: Fix out-of-bounds access (#7085 )	2023-10-22 11:07:06 -07:00
Castor215	4ac10c4a9d	externals: allow users to use system Zstandard (#7083 )	2023-10-21 16:10:02 -07:00
Castor215	2416258117	externals: add overarching USE_SYSTEM_LIBS variable (#7078 )	2023-10-20 17:02:20 -07:00
Steveice10	6d4e462e42	build: Ensure we default to Release build type. (#7080 )	2023-10-19 15:02:49 -07:00
GPUCode	ef43776c7b	shader: Fix address register offset behavior in x64 Jit (#6942 ) * shader: Fix address register offset behavior in x64 Jit * shader: Remove redundant jump * tests: Add address register tests * shader: Remove additional pre-multiplications by 16 * tests: Add catch-stringifier for vec4f * tests: Format	2023-10-18 19:41:36 +03:00