715 changed files with 33973 additions and 191784 deletions
--- a/.gitattributes
+++ b/.gitattributes
@ -1,18 +0,0 @@
-# Vendored Dependencies
-src/frontend/glad/**                         linguist-vendored
-src/frontend/qt_sdl/gif-h/**                 linguist-vendored
-src/frontend/qt_sdl/toml/**                  linguist-vendored
-src/net/libslirp/**                          linguist-vendored
-src/net/pcap/**                              linguist-vendored
-src/sha1/**                                  linguist-vendored
-src/teakra/**                                linguist-vendored
-src/tiny-AES-c/**                            linguist-vendored
-src/xxhash/**                                linguist-vendored
-
-# A handful of custom files embedded in the vendored dependencies
-
-## Ad-hoc CMakeLists.txt for melonDS
-src/net/libslirp/src/CMakeLists.txt          -linguist-vendored
-
-## glib stub
-src/net/libslirp/src/glib/**                 -linguist-vendored
--- a/.github/FUNDING.yml
+++ b/.github/FUNDING.yml
@ -1,2 +1 @@
-patreon: Arisotura
-custom: ["https://paypal.me/Arisotura", "http://melonds.kuribo64.net/donate.php"]
+patreon: staplebutter
--- a/.github/workflows/build-macos.yml
+++ b/.github/workflows/build-macos.yml
@ -1,94 +1,39 @@
-name: macOS
+name: CMake Build (macOS x86-64)

 on:
  push:
    branches:
      - master
-      - ci/*
  pull_request:
    branches:
      - master

 env:
-  VCPKG_COMMIT: 2ad004460f5db4d3b66f62f5799ff66c265c4b5d
-  MELONDS_GIT_BRANCH: ${{ github.ref }}
-  MELONDS_GIT_HASH: ${{ github.sha }}
-  MELONDS_BUILD_PROVIDER: GitHub Actions
-  MELONDS_VERSION_SUFFIX: " RC"
+  BUILD_TYPE: Release

 jobs:
-  build-macos:
-    strategy:
-      matrix:
-        arch: [x86_64, arm64]
+  build:

-    name: ${{ matrix.arch }}
-    runs-on: macos-14
-    steps:
-      - name: Check out sources
-        uses: actions/checkout@v3
-      - name: Install dependencies for package building
-        run: |
-          brew install autoconf automake autoconf-archive libtool python-setuptools
-      - name: Set up CMake
-        uses: lukka/get-cmake@latest
-      - name: Set up vcpkg
-        uses: lukka/run-vcpkg@v11
-        with:
-          vcpkgGitCommitId: ${{ env.VCPKG_COMMIT }}
-      - name: Build
-        uses: lukka/run-cmake@v10
-        with:
-          configurePreset: release-mac-${{ matrix.arch }}
-          buildPreset: release-mac-${{ matrix.arch }}
-          configurePresetAdditionalArgs: "['-DMELONDS_EMBED_BUILD_INFO=ON']"
-      - name: Compress app bundle
-        shell: bash
-        run: |
-          cd build/release-mac-${{ matrix.arch }}
-          zip -r -y ../../macOS-${{ matrix.arch }}.zip melonDS.app
-      - name: Upload artifact
-        uses: actions/upload-artifact@v4
-        with:
-          name: macOS-${{ matrix.arch }}
-          path: macOS-${{ matrix.arch }}.zip
-          retention-days: 1
+    runs-on: macos-latest

-  universal-binary:
-    name: Universal binary
-    needs: [build-macos]
-    runs-on: macos-13
-    continue-on-error: true
    steps:
-      - name: Download x86_64
-        uses: actions/download-artifact@v4
-        with:
-          name: macOS-x86_64
-          path: x86_64
-      - name: Download arm64
-        uses: actions/download-artifact@v4
-        with:
-          name: macOS-arm64
-          path: arm64
-      - name: Combine app bundles
-        shell: bash
-        run: |
-          unzip x86_64/*.zip -d x86_64
-          unzip arm64/*.zip -d arm64
-          lipo {x86_64,arm64}/melonDS.app/Contents/MacOS/melonDS -create -output melonDS
-          cp -a arm64/melonDS.app melonDS.app
-          cp melonDS melonDS.app/Contents/MacOS/melonDS
-          codesign -s - --deep melonDS.app
-          zip -r -y macOS-universal.zip melonDS.app
-      - name: Upload artifact
-        uses: actions/upload-artifact@v4
-        with:
-          name: macOS-universal
-          path: macOS-universal.zip
-#     - name: Clean up architecture-specific artifacts
-#       uses: geekyeggo/delete-artifact@v4
-#       with:
-#         failOnError: false
-#         name: |
-#           macOS-x86_64
-#           macOS-arm64
+    - uses: actions/checkout@v1
+    - name: Install dependencies
+      working-directory: ${{runner.workspace}}
+      run: |
+        brew install cmake sdl2 qt5 libslirp
+    - name: Create build environment
+      run: mkdir ${{runner.workspace}}/build
+    - name: Configure
+      working-directory: ${{runner.workspace}}/build
+      run: cmake $GITHUB_WORKSPACE -DCMAKE_BUILD_TYPE=$BUILD_TYPE -DQt5_DIR=$(brew --prefix qt5)/lib/cmake/Qt5
+    - name: Make
+      working-directory: ${{runner.workspace}}/build
+      run: |
+        make -j$(sysctl -n hw.ncpu)
+        mkdir dist
+        cp -r melonDS.app dist
+    - uses: actions/upload-artifact@v1
+      with:
+        name: melonDS.app
+        path: ${{runner.workspace}}/build/dist
--- a/.github/workflows/build-ubuntu-aarch64.yml
+++ b/.github/workflows/build-ubuntu-aarch64.yml
@ -0,0 +1,57 @@
+name: CMake Build (Ubuntu aarch64)
+
+on:
+  push:
+    branches:
+      - master
+  pull_request:
+    branches:
+      - master
+
+env:
+  BUILD_TYPE: Release
+
+jobs:
+  build:
+    runs-on: ubuntu-20.04
+
+    steps:
+    - uses: actions/checkout@v1
+    - name: Upgrade system
+      shell: bash
+      working-directory: ${{runner.workspace}}
+      run: | #Fix grub installation error - https://github.com/actions/virtual-environments/issues/1605
+        sudo apt update
+        sudo apt-get install grub-efi  
+        sudo update-grub
+        sudo apt full-upgrade
+    - name: Install dependencies
+      shell: bash
+      working-directory: ${{runner.workspace}}
+      run: |
+        sudo rm -f /etc/apt/sources.list.d/kubernetes.list
+        sudo dpkg --add-architecture arm64
+        sudo sh -c "sed \"s|^deb \([a-z\.:/]*\) \([a-z\-]*\) \(.*\)$|deb [arch=amd64] \1 \2 \3\ndeb [arch=arm64] http://ports.ubuntu.com/ubuntu-ports \2 \3|\" /etc/apt/sources.list > /etc/apt/sources.list.new"
+        sudo rm /etc/apt/sources.list
+        sudo mv /etc/apt/sources.list{.new,}
+        sudo apt update
+        sudo apt install aptitude
+        sudo aptitude install -y {gcc-10,g++-10,pkg-config}-aarch64-linux-gnu libsdl2-dev:arm64 qtbase5-dev:arm64 libslirp-dev:arm64
+    - name: Create build environment
+      run: mkdir ${{runner.workspace}}/build
+    - name: Configure
+      shell: bash
+      working-directory: ${{runner.workspace}}/build
+      run: |
+          CC=aarch64-linux-gnu-gcc-10 CXX=aarch64-linux-gnu-g++-10 cmake -DPKG_CONFIG_EXECUTABLE=/usr/bin/aarch64-linux-gnu-pkg-config $GITHUB_WORKSPACE -DCMAKE_BUILD_TYPE=$BUILD_TYPE
+    - name: Make
+      shell: bash
+      working-directory: ${{runner.workspace}}/build
+      run: |
+        make -j$(nproc --all) \
+        && mkdir dist \
+        && cp melonDS dist
+    - uses: actions/upload-artifact@v1
+      with:
+        name: melonDS
+        path: ${{runner.workspace}}/build/dist
--- a/.github/workflows/build-ubuntu.yml
+++ b/.github/workflows/build-ubuntu.yml
@ -1,63 +1,37 @@
-name: Ubuntu
+name: CMake Build (Ubuntu x86-64)

 on:
  push:
    branches:
      - master
-      - ci/*
  pull_request:
    branches:
      - master

-env:
-  MELONDS_GIT_BRANCH: ${{ github.ref }}
-  MELONDS_GIT_HASH: ${{ github.sha }}
-  MELONDS_BUILD_PROVIDER: GitHub Actions
-  MELONDS_VERSION_SUFFIX: " RC"
-
 jobs:
  build:
-    continue-on-error: true
-    strategy:
-      matrix:
-        arch:
-          - runner: ubuntu-22.04
-            name: x86_64
-          - runner: ubuntu-22.04-arm
-            name: aarch64

-    name: ${{ matrix.arch.name }}
-    runs-on: ${{ matrix.arch.runner }}
+    runs-on: ubuntu-20.04

    steps:
-    - uses: actions/checkout@v4
-      name: Check out sources
+    - uses: actions/checkout@v1
    - name: Install dependencies
      run: |
+        sudo rm -f /etc/apt/sources.list.d/dotnetdev.list /etc/apt/sources.list.d/microsoft-prod.list
        sudo apt update
-        sudo apt install --allow-downgrades cmake ninja-build extra-cmake-modules libpcap0.8-dev libsdl2-dev libenet-dev \
-          qt6-{base,base-private,multimedia}-dev libqt6svg6-dev libarchive-dev libzstd-dev libfuse2
+        sudo apt install cmake libcurl4-gnutls-dev libpcap0.8-dev libsdl2-dev qt5-default libslirp0 libslirp-dev --allow-downgrades
+    - name: Create build environment
+      run: mkdir ${{runner.workspace}}/build
    - name: Configure
-      run: cmake -B build -G Ninja -DCMAKE_INSTALL_PREFIX=/usr -DMELONDS_EMBED_BUILD_INFO=ON
-    - name: Build
+      working-directory: ${{runner.workspace}}/build
+      run: cmake $GITHUB_WORKSPACE
+    - name: Make
+      working-directory: ${{runner.workspace}}/build
      run: |
-        cmake --build build
-        DESTDIR=AppDir cmake --install build
-    - uses: actions/upload-artifact@v4
+        make -j$(nproc --all)
+        mkdir dist
+        cp melonDS dist
+    - uses: actions/upload-artifact@v1
      with:
-        name: melonDS-ubuntu-${{ matrix.arch.name }}
-        path: AppDir/usr/bin/melonDS
-    - name: Fetch AppImage tools
-      run: |
-        wget https://github.com/linuxdeploy/linuxdeploy/releases/download/continuous/linuxdeploy-${{ matrix.arch.name }}.AppImage
-        wget https://github.com/linuxdeploy/linuxdeploy-plugin-qt/releases/download/continuous/linuxdeploy-plugin-qt-${{ matrix.arch.name }}.AppImage
-        chmod a+x linuxdeploy-*.AppImage
-    - name: Build the AppImage
-      env:
-        QMAKE: /usr/lib/qt6/bin/qmake
-      run: |
-        ./linuxdeploy-${{ matrix.arch.name }}.AppImage --appdir AppDir --plugin qt --output appimage
-    - uses: actions/upload-artifact@v4
-      with:
-        name: melonDS-appimage-${{ matrix.arch.name }}
-        path: melonDS*.AppImage
+        name: melonDS
+        path: ${{runner.workspace}}/build/dist
--- a/.github/workflows/build-windows.yml
+++ b/.github/workflows/build-windows.yml
@ -1,45 +1,43 @@
-name: Windows
+name: CMake Build (Windows x86-64)

 on:
  push:
    branches:
      - master
-      - ci/*
  pull_request:
    branches:
      - master

 env:
-  VCPKG_COMMIT: 2ad004460f5db4d3b66f62f5799ff66c265c4b5d
-  MELONDS_GIT_BRANCH: ${{ github.ref }}
-  MELONDS_GIT_HASH: ${{ github.sha }}
-  MELONDS_BUILD_PROVIDER: GitHub Actions
-  MELONDS_VERSION_SUFFIX: " RC"
+  BUILD_TYPE: Release

 jobs:
  build:
+
    runs-on: windows-latest
-    defaults:
-      run:
-        shell: msys2 {0}
+
    steps:
-    - name: Check out sources
-      uses: actions/checkout@v3
-    - name: Set up MSYS2
-      uses: msys2/setup-msys2@v2
-      with:
-        msystem: ucrt64
-        update: true
-        pacboy: gcc:p cmake:p ninja:p make:p
-    - name: Set up vcpkg
-      uses: lukka/run-vcpkg@v11
-      with:
-        vcpkgGitCommitId: ${{ env.VCPKG_COMMIT }}
+    - uses: actions/checkout@v1
+    - name: Install MSYS2
+      working-directory: ${{runner.workspace}}
+      run: | # Fetch MSYS2 build from XQEmu. Official distribution causes a CI failure due to permission errors.
+        choco install msys2
+        C:\tools\msys64\usr\bin\bash.exe -lc "pacman -Syuq --noconfirm"
+    - name: Install dependencies
+      run: C:\tools\msys64\usr\bin\bash.exe -lc "pacman -Sq --noconfirm git make mingw-w64-x86_64-{cmake,mesa,SDL2,qt5-static,libslirp,toolchain}"
+    - name: Create build environment
+      run: |
+        New-Item -ItemType directory -Path ${{runner.workspace}}\melonDS\build
+        Copy-Item -Path ${{runner.workspace}}\melonDS -Destination C:\tools\msys64\home\runneradmin -Recurse
    - name: Configure
-      run: cmake --preset=release-mingw-x86_64 -DMELONDS_EMBED_BUILD_INFO=ON
-    - name: Build
-      run: cmake --build --preset=release-mingw-x86_64
-    - uses: actions/upload-artifact@v4
+      run: |
+        C:\tools\msys64\usr\bin\bash.exe -lc "export PATH=`"/mingw64/bin:`$PATH`" \
+        && cd melonDS/build && cmake .. -G 'MSYS Makefiles' -DCMAKE_BUILD_TYPE=${{env.BUILD_TYPE}} -DBUILD_STATIC=ON -DQT5_STATIC_DIR=C:/tools/msys64/mingw64/qt5-static"
+    - name: Make
+      run: |
+        C:\tools\msys64\usr\bin\bash.exe -lc "export PATH=`"/mingw64/bin:`$PATH`" \
+        && cd melonDS/build && make -j$(nproc --all)"
+    - uses: actions/upload-artifact@v1
      with:
-        name: melonDS-windows-x86_64
-        path: .\build\release-mingw-x86_64\melonDS.exe
+        name: melonDS
+        path: C:\tools\msys64\home\runneradmin\melonDS\build\melonDS.exe
--- a/.gitignore
+++ b/.gitignore
@ -1,4 +1,4 @@
-build*/
+build
 bin
 obj
 *.depend
@ -6,18 +6,10 @@ obj
 *.o
 melon_grc.c
 melon_grc.h
-melon.rc
-cmake-build*
+cmake-build
 cmake-build-debug
-compile_commands.json
 .idea
-.cache

 *.exe

 .DS_Store
-
-.vs
-.vscode
-CMakeFiles
-CMakeCache.txt
--- a/BUILD.md
+++ b/BUILD.md
@ -1,81 +0,0 @@
-# Building melonDS
-
-* [Linux](#linux)
-* [Windows](#windows)
-* [macOS](#macos)
-
-## Linux
-1. Install dependencies:
-   * Ubuntu:
-     * All versions: `sudo apt install cmake extra-cmake-modules libcurl4-gnutls-dev libpcap0.8-dev libsdl2-dev libarchive-dev libenet-dev libzstd-dev`
-     * 24.04: `sudo apt install qt6-{base,base-private,multimedia,svg}-dev`
-     * 22.04: `sudo apt install qtbase6-dev qtbase6-private-dev qtmultimedia6-dev libqt6svg6-dev`
-     * Older versions: `sudo apt install qtbase5-dev qtbase5-private-dev qtmultimedia5-dev libqt5svg5-dev`  
-       Also add `-DUSE_QT6=OFF` to the first CMake command below.
-   * Fedora: `sudo dnf install gcc-c++ cmake extra-cmake-modules SDL2-devel libarchive-devel enet-devel libzstd-devel qt6-{qtbase,qtbase-private,qtmultimedia,qtsvg}-devel wayland-devel`
-   * Arch Linux: `sudo pacman -S base-devel cmake extra-cmake-modules git libpcap sdl2 qt6-{base,multimedia,svg} libarchive enet zstd`
-2. Download the melonDS repository and prepare:
-   ```bash
-   git clone https://github.com/melonDS-emu/melonDS
-   cd melonDS
-   ```
-3. Compile:
-   ```bash
-   cmake -B build
-   cmake --build build -j$(nproc --all)
-   ```
-
-## Windows
-1. Install [MSYS2](https://www.msys2.org/)
-2. Open the MSYS2 terminal from the Start menu:
-   * For x64 systems (most common), use **MSYS2 UCRT64**
-   * For ARM64 systems, use **MSYS2 CLANGARM64**
-3. Update the packages using `pacman -Syu` and reopen the same terminal if it asks you to
-4. Install git and clone the repository
-   ```bash
-   pacman -S git
-   git clone https://github.com/melonDS-emu/melonDS
-   cd melonDS
-   ```
-5. Install dependencies:  
-   Replace `<prefix>` below with `mingw-w64-ucrt-x86_64` on x64 systems, or `mingw-w64-clang-aarch64` on ARM64 systems.
-   ```bash
-   pacman -S <prefix>-{toolchain,cmake,SDL2,libarchive,enet,zstd}
-   ```
-6. Install Qt and configure the build directory
-   * Dynamic builds (with DLLs)
-     1. Install Qt: `pacman -S <prefix>-{qt6-base,qt6-svg,qt6-multimedia,qt6-svg,qt6-tools}`
-     2. Set up the build directory with `cmake -B build`
-   * Static builds (without DLLs, standalone executable)
-     1. Install Qt: `pacman -S <prefix>-qt5-static`  
-        (Note: As of writing, the `qt6-static` package does not work.)
-     2. Set up the build directory with `cmake -B build -DBUILD_STATIC=ON -DUSE_QT6=OFF -DCMAKE_PREFIX_PATH=$MSYSTEM_PREFIX/qt5-static`
-7. Compile: `cmake --build build`
-
-If everything went well, melonDS should now be in the `build` folder. For dynamic builds, you may need to run melonDS from the MSYS2 terminal in order for it to find the required DLLs.
-
-## macOS
-1. Install the [Homebrew Package Manager](https://brew.sh)
-2. Install dependencies: `brew install git pkg-config cmake sdl2 qt@6 libarchive enet zstd`
-3. Download the melonDS repository and prepare:
-   ```zsh
-   git clone https://github.com/melonDS-emu/melonDS
-   cd melonDS
-   ```
-4. Compile:
-   ```zsh
-   cmake -B build -DCMAKE_PREFIX_PATH="$(brew --prefix qt@6);$(brew --prefix libarchive)"
-   cmake --build build -j$(sysctl -n hw.logicalcpu)
-   ```
-If everything went well, melonDS.app should now be in the `build` directory.
-
-### Self-contained app bundle
-If you want an app bundle that can be distributed to other computers without needing to install dependencies through Homebrew, you can additionally run `
-../tools/mac-libs.rb .` after the build is completed, or add `-DMACOS_BUNDLE_LIBS=ON` to the first CMake command.
-
-## Nix (macOS/Linux)
-
-melonDS provides a Nix flake with support for both macOS and Linux. The [Nix package manager](https://nixos.org) needs to be installed to use it.
-
-* To run melonDS, just type `nix run github:melonDS-emu/melonDS`.
-* To get a shell for development, clone the melonDS repository and type `nix develop` in its directory.
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@ -1,44 +1,28 @@
-cmake_minimum_required(VERSION 3.16)
+cmake_minimum_required(VERSION 3.13)

-cmake_policy(VERSION 3.15)
+include(CheckSymbolExists)
+include(CheckLibraryExists)
+
+cmake_policy(VERSION 3.13)
 if (POLICY CMP0076)
    cmake_policy(SET CMP0076 NEW)
 endif()

-set(CMAKE_POLICY_DEFAULT_CMP0069 NEW)
+set(CMAKE_OSX_DEPLOYMENT_TARGET "10.9" CACHE STRING "Minimum OS X deployment version")

-set(CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/cmake" ${CMAKE_MODULE_PATH})
-set(CMAKE_USER_MAKE_RULES_OVERRIDE "${CMAKE_CURRENT_SOURCE_DIR}/cmake/DefaultBuildFlags.cmake")
-set(CMAKE_OSX_DEPLOYMENT_TARGET "10.15" CACHE STRING "Minimum OS X deployment version")
-
-option(USE_VCPKG "Use vcpkg for dependency packages" OFF)
-if (USE_VCPKG)
-	include(ConfigureVcpkg)
-endif()
-
-project(melonDS
-    VERSION 1.0
-    DESCRIPTION "DS emulator, sorta"
-    HOMEPAGE_URL "https://melonds.kuribo64.net"
-    LANGUAGES C CXX)
-
-include(CheckSymbolExists)
-include(CheckLibraryExists)
-include(CMakeDependentOption)
-include(CheckIPOSupported)
-
-include(SetupCCache)
-include(Sanitizers)
+project(melonDS CXX)

 set(CMAKE_C_STANDARD 11)
-set(CMAKE_C_STANDARD_REQUIRED ON)
 set(CMAKE_CXX_STANDARD 17)
-set(CMAKE_CXX_STANDARD_REQUIRED ON)
 set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR})

-if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
-    set(CMAKE_BUILD_TYPE Release CACHE STRING "Choose the type of build." FORCE)
-    set_property(CACHE CMAKE_BUILD_TYPE PROPERTY STRINGS "Debug" "Release" "MinSizeRel" "RelWithDebInfo")
+check_library_exists(m pow "" LIBM)
+if(LIBM)
+    link_libraries(m)
+endif()
+
+if (NOT CMAKE_BUILD_TYPE)
+    set(CMAKE_BUILD_TYPE Release)
 endif()

 function(detect_architecture symbol arch)
@ -62,40 +46,59 @@ detect_architecture("__i386__" x86)
 detect_architecture("__arm__" ARM)
 detect_architecture("__aarch64__" ARM64)

-cmake_dependent_option(ENABLE_JIT "Enable JIT recompiler" ON
-    "ARCHITECTURE STREQUAL x86_64 OR ARCHITECTURE STREQUAL ARM64" OFF)
-cmake_dependent_option(ENABLE_JIT_PROFILING "Enable JIT profiling with VTune" OFF "ENABLE_JIT" OFF)
+if (ARCHITECTURE STREQUAL x86_64 OR ARCHITECTURE STREQUAL ARM64)
+	option(ENABLE_JIT "Enable x64 JIT recompiler" ON)
+endif()
+
+if (ENABLE_JIT)
+	add_definitions(-DJIT_ENABLED)
+endif()
+
+if (CMAKE_BUILD_TYPE STREQUAL Release)
+	option(ENABLE_LTO "Enable link-time optimization" ON)
+else()
+	option(ENABLE_LTO "Enable link-time optimization" OFF)
+endif()
+
 option(ENABLE_OGLRENDERER "Enable OpenGL renderer" ON)

-check_ipo_supported(RESULT IPO_SUPPORTED)
-cmake_dependent_option(ENABLE_LTO_RELEASE "Enable link-time optimizations for release builds" ON "IPO_SUPPORTED" OFF)
-cmake_dependent_option(ENABLE_LTO "Enable link-time optimizations" OFF "IPO_SUPPORTED" OFF)
-
-if (ENABLE_LTO_RELEASE)
-    set(CMAKE_INTERPROCEDURAL_OPTIMIZATION_RELEASE TRUE)
+if (ENABLE_OGLRENDERER)
+    add_definitions(-DOGLRENDERER_ENABLED)
 endif()

-if (ENABLE_LTO)
-    set(CMAKE_INTERPROCEDURAL_OPTIMIZATION TRUE)
+if (CMAKE_BUILD_TYPE STREQUAL Debug)
+	add_compile_options(-Og)
 endif()

-if (NOT APPLE)
-    set(CMAKE_EXE_LINKER_FLAGS_RELEASE "${CMAKE_EXE_LINKER_FLAGS_RELEASE} -s")
+if (CMAKE_BUILD_TYPE STREQUAL Release)
+	add_compile_options(-O3)
+	add_link_options(-s)
 endif()

 if (WIN32)
-    option(BUILD_STATIC "Statically link dependencies" OFF)
+	option(BUILD_STATIC "Statically link dependencies" OFF)
 endif()

-if (BUILD_STATIC AND WIN32)
-    set(CMAKE_FIND_LIBRARY_SUFFIXES .a)
+if (ENABLE_LTO)
+	if (WIN32)
+		add_compile_options(-flto)
+		add_link_options(-flto)
+	else()
+		add_compile_options(-flto -fPIC)
+		add_link_options(-flto -fuse-linker-plugin -pie)
+	endif()
 endif()

-set(CMAKE_POSITION_INDEPENDENT_CODE ON)
-
-option(ENABLE_GDBSTUB "Enable GDB stub" ON)
-if (ENABLE_GDBSTUB)
-    add_definitions(-DGDBSTUB_ENABLED)
+if (CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
+	set(CMAKE_AR "gcc-ar")
+	set(CMAKE_RANLIB "gcc-ranlib")
+elseif (CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
+	find_program(LLD NAMES ld.lld ld64.lld lld-link)
+	if (NOT LLD STREQUAL "LLD-NOTFOUND")
+		add_link_options(-fuse-ld=lld)
+	endif()
+	set(CMAKE_AR "llvm-ar")
+	set(CMAKE_RANLIB "llvm-ranlib")
 endif()

 option(BUILD_QT_SDL "Build Qt/SDL frontend" ON)
@ -103,5 +106,5 @@ option(BUILD_QT_SDL "Build Qt/SDL frontend" ON)
 add_subdirectory(src)

 if (BUILD_QT_SDL)
-    add_subdirectory(src/frontend/qt_sdl)
-endif()
+	add_subdirectory(src/frontend/qt_sdl)
+endif()
--- a/CMakePresets.json
+++ b/CMakePresets.json
@ -1,105 +0,0 @@
-{
-  "version": 6,
-  "configurePresets": [
-    {
-      "name": "release",
-      "displayName": "Release",
-      "description": "Default release build configuration.",
-      "generator": "Ninja",
-      "binaryDir": "${sourceDir}/build/release"
-    },
-    {
-      "inherits": "release",
-      "name": "release-vcpkg",
-      "displayName": "Release (vcpkg)",
-      "description": "Release build with packages from vcpkg.",
-      "cacheVariables": {
-        "USE_VCPKG": {
-          "type": "BOOL",
-          "value": "ON"
-        }
-      }
-    },
-    {
-      "name": "release-mingw-x86_64",
-      "inherits": "release-vcpkg",
-      "displayName": "Windows MinGW release (x86_64)",
-      "binaryDir": "${sourceDir}/build/release-mingw-x86_64",
-      "generator": "Ninja",
-      "cacheVariables": {
-        "BUILD_STATIC": {
-          "type": "BOOL",
-          "value": "ON"
-        }
-      }
-    },
-    {
-      "name": "release-mac-x86_64",
-      "inherits": "release-vcpkg",
-      "displayName": "macOS release (x86_64)",
-      "binaryDir": "${sourceDir}/build/release-mac-x86_64",
-      "cacheVariables": { "CMAKE_OSX_ARCHITECTURES": "x86_64" }
-    },
-    {
-      "name": "release-mac-arm64",
-      "inherits": "release-vcpkg",
-      "displayName": "macOS release (arm64)",
-      "binaryDir": "${sourceDir}/build/release-mac-arm64",
-      "cacheVariables": { "CMAKE_OSX_ARCHITECTURES": "arm64" }
-    }
-  ],
-  "buildPresets": [
-    {
-      "name": "release",
-      "configurePreset": "release"
-    },
-    {
-      "name": "release-vcpkg",
-      "configurePreset": "release-vcpkg"
-    },
-    {
-      "name": "release-mingw-x86_64",
-      "configurePreset": "release-mingw-x86_64"
-    },
-    {
-      "name": "release-mac-x86_64",
-      "configurePreset": "release-mac-x86_64"
-    },
-    {
-      "name": "release-mac-arm64",
-      "configurePreset": "release-mac-arm64"
-    }
-  ],
-  "workflowPresets": [
-    {
-      "name": "release",
-      "displayName": "Release",
-      "steps": [
-        { "type": "configure", "name": "release" },
-        { "type": "build", "name": "release" }
-      ]
-    },
-    {
-      "name": "release-vcpkg",
-      "displayName": "Release (vcpkg)",
-      "steps": [
-        { "type": "configure", "name": "release-vcpkg" },
-        { "type": "build", "name": "release-vcpkg" }
-      ]
-    },
-    {
-      "name": "release-mac-x86_64",
-      "steps": [
-        { "type": "configure", "name": "release-mac-x86_64" },
-        { "type": "build", "name": "release-mac-x86_64" }
-      ]
-    },
-    {
-      "name": "release-mac-arm64",
-      "steps": [
-        { "type": "configure", "name": "release-mac-arm64" },
-        { "type": "build", "name": "release-mac-arm64" }
-      ]
-    }
-  ]
-}
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@ -1,150 +0,0 @@
-# Contributor guide for melonDS
-
-Please follow a style as documented here. Note that this guide was not always enforced, so some parts of the code violate it.
-
-Files should use 4 spaces for indentation.
-
-```c++
-
-// for single line comments prefer C++ style
-
-/*
-    for multiline comments
-    both C style comments
-*/
-// as well as
-// C++ style comments are possible
-
-// when including headers from the C standard library use their C names (so don't use cstdio/cstring, ...)
-#include <stdio.h>
-
-// namespaces in PascalCase
-namespace Component
-{ // for all constructs curly braces go onto the following line
-
-// the content of namespaces should not be indented
-
-int GlobalVariable; // in PascalCase
-
-// function names should use PascalCase, parameters camelCase:
-void Test(int someParam)
-{
-    int variable = someParam * 2; // local variables in camelCase
-    
-    // you can slightly vary the spacing around operators:
-    int variable2 = someParam*2 + 1;
-    // but avoid something like this: someParam* 2+ 3
-
-    for (int i = 0; i < variable; i++) // always a space between if/for/while and the braces
-    {
-        // not using curly braces is allowed
-        // if the body of the if/for/while is simple:
-        if ((i % 2) == 0)
-            printf("%d\n", i); // no space between the function name and the braces
-    }
-}
-
-// defines should always be in CAPITALISED_SNAKE_CASE
-#ifdef SOME_CONFIGURATION
-
-// the content of #ifdef sections should not be indented
-// the only exception being otherwise hard to read nested sections of #ifdef/#if/#defines
-// as e.g. in ARMJIT_Memory.cpp
-
-// prefer #ifdef/#ifndef, only use if defined(...) for complex expressions like this:
-#if defined(THIS) || defined(THAT)
-// ...
-#endif
-
-class MyClass // PascalCase
-{
-public: // access specfications are not indented
-    void Test(int param) // for methods the same rules apply as for functions
-    {
-    }
-
-private:
-    int MemberVariable; // PascalCase, no prefix
-};
-
-#endif
-
-#endif
-
-enum
-{
-    // enums should always have a common prefix in camelCase
-    // separated by an underscore with the item name
-    // which has to be in PascalCase
-    enumPrefix_FirstElement,
-    enumPrefix_SecondElement,
-    enumPrefix_ThirdElement,
-    enumPrefix_FourthElement,
-};
-
-}
-
-```
-
-Some additional notes:
-
-* Keep the definition and initialisation of local variables in one place and keep the scope of local variables as small as possible.
-
-**That means avoid code like this**:
-```cpp
-void ColorConvert(u32* dst, u16* vram)
-{
-    u16 color;
-    u8 r, g, b;
-    int i;
-
-    for (i = 0; i < 256; i++)
-    {
-        color = vram[i];
-        r = (color & 0x001F) << 1;
-        g = (color & 0x03E0) >> 4;
-        b = (color & 0x7C00) >> 9;
-
-        dst[i] = r | (g << 8) | (b << 16);
-    }
-}
-```
-
-**Do this instead:**
-```cpp
-void ColorConvert(u32* dst, u16* vram)
-{
-    for (int i = 0; i < 256; i++)
-    {
-        u16 color = vram[i];
-        u8 r = (color & 0x001F) << 1;
-        u8 g = (color & 0x03E0) >> 4;
-        u8 b = (color & 0x7C00) >> 9;
-
-        dst[i] = r | (g << 8) | (b << 16);
-    }
-}
-```
-
-* For integer types preferably use the explictly typed ones. We have short aliases for them defined in types.h (for unsigned types: `u8`, `u16`, `u32`, `u16`. For signed `s8`, `s16`, `s32`, `s64`). In some situations like loop variables, using `int` is possible as well.
-* Don't overdo object oriented programming. Always try to use a simpler construct first, only use a polymorphic class if a namespace with functions in it doesn't cut it.
-
-* In doubt put a namespace around your part of the code.
-
-* C style strings (and the associated functions from the C standard library) are used in most places. We are thinking about changing this, as C strings are a bit of a hassle to deal with, but for the time being this is what we use.
-
-* Only the C standard IO is used (so use `printf`, `fopen`, … Do not use `std::cout`/`std::ostream`, …).
-
-* Complex C++ containers can be used (`std::vector`, `std::list`, `std::map`, …). `std::array` is usually not used, unless necessary so that the container can be used with other C++ constructs (e.g. `<algorithm>`). Only use them if a C array doesn't cut it.
-
-* File names should be in PascalCase
-
-* If a header file is called MyHeader.h it should be guarded with an ifdef like this:
-```cpp
-#ifndef MYHEADER_H
-#define MYHEADER_H
-// ...
-#endif
-```
-
-* And at last, if you have any questions, visit us on IRC (see the readme)!
--- a/README.md
+++ b/README.md
@ -1,15 +1,10 @@
-<p align="center"><img src="https://raw.githubusercontent.com/melonDS-emu/melonDS/master/res/icon/melon_128x128.png"></p>
+<p align="center"><img src="https://raw.githubusercontent.com/StapleButter/melonDS/master/icon/melon_128x128.png"></p>
 <h2 align="center"><b>melonDS</b></h2>
 <p align="center">
 <a href="http://melonds.kuribo64.net/" alt="melonDS website"><img src="https://img.shields.io/badge/website-melonds.kuribo64.net-%2331352e.svg"></a>
-<a href="http://melonds.kuribo64.net/downloads.php" alt="Release: 0.9.5"><img src="https://img.shields.io/badge/release-0.9.5-%235c913b.svg"></a>
+<a href="http://melonds.kuribo64.net/downloads.php" alt="Release: 0.9.1"><img src="https://img.shields.io/badge/release-0.9.1-%235c913b.svg"></a>
 <a href="https://www.gnu.org/licenses/gpl-3.0" alt="License: GPLv3"><img src="https://img.shields.io/badge/License-GPL%20v3-%23ff554d.svg"></a>
 <a href="https://kiwiirc.com/client/irc.badnik.net/?nick=IRC-Source_?#melonds" alt="IRC channel: #melonds"><img src="https://img.shields.io/badge/IRC%20chat-%23melonds-%23dd2e44.svg"></a>
-<a href="https://discord.gg/pAMAtExcqV" alt="Discord"><img src="https://img.shields.io/badge/Discord-Kuribo64-7289da?logo=discord&logoColor=white"></a>
-<br>
-<a href="https://github.com/melonDS-emu/melonDS/actions/workflows/build-windows.yml?query=event%3Apush"><img src="https://github.com/melonDS-emu/melonDS/actions/workflows/build-windows.yml/badge.svg" /></a>
-<a href="https://github.com/melonDS-emu/melonDS/actions/workflows/build-ubuntu.yml?query=event%3Apush"><img src="https://github.com/melonDS-emu/melonDS/actions/workflows/build-ubuntu.yml/badge.svg" /></a>
-<a href="https://github.com/melonDS-emu/melonDS/actions/workflows/build-macos.yml?query=event%3Apush"><img src="https://github.com/melonDS-emu/melonDS/actions/workflows/build-macos.yml/badge.svg" /></a>
 </p>
 DS emulator, sorta

@ -18,7 +13,12 @@ The goal is to do things right and fast, akin to blargSNES (but hopefully better

 ## How to use

-Firmware boot (not direct boot) requires a BIOS/firmware dump from an original DS or DS Lite.
+melonDS requires BIOS/firmware copies from a DS. Files required:
+ * bios7.bin, 16KB: ARM7 BIOS
+ * bios9.bin, 4KB: ARM9 BIOS
+ * firmware.bin, 128/256/512KB: firmware
+
+Firmware boot requires a firmware dump from an original DS or DS Lite.
 DS firmwares dumped from a DSi or 3DS aren't bootable and only contain configuration data, thus they are only suitable when booting games directly.

 ### Possible firmware sizes
@ -32,20 +32,86 @@ DS BIOS dumps from a DSi or 3DS can be used with no compatibility issues. DSi BI
 As for the rest, the interface should be pretty straightforward. If you have a question, don't hesitate to ask, though!

 ## How to build
-See [BUILD.md](./BUILD.md) for build instructions.

+### Linux:
+
+* Install dependencies:
+
+```sh
+sudo apt-get install cmake libcurl4-gnutls-dev libpcap0.8-dev libsdl2-dev qtbase5-dev qtdeclarative5-dev libslirp-dev
+```
+
+* Compile:
+
+```sh
+mkdir -p build
+cd build
+cmake ..
+make -j$(nproc --all)
+```
+
+### Windows:
+
+1. Install [MSYS2](https://www.msys2.org/)
+2. Open the **MSYS2 MinGW 64-bit** terminal
+3. Update the packages using `pacman -Syu` and reopen the terminal if it asks you to
+
+#### Dynamic builds (with DLLs)
+4. Install dependencies: `pacman -S git make mingw-w64-x86_64-{cmake,mesa,SDL2,toolchain,qt5,libslirp}`
+5. Run the following commands
+   ```bash
+   git clone https://github.com/Arisotura/melonDS.git
+   cd melonDS
+   mkdir build
+   cd build
+   cmake .. -G "MSYS Makefiles"
+   make -j$(nproc --all)
+   ../msys-dist.sh
+   ```
+If everything went well, melonDS and the libraries it needs should now be in the `dist` folder.
+
+#### Static builds (without DLLs, standalone executable)
+4. Install dependencies: `pacman -S git make mingw-w64-x86_64-{cmake,mesa,SDL2,toolchain,qt5-static,libslirp}`
+5. Run the following commands
+   ```bash
+   git clone https://github.com/Arisotura/melonDS.git
+   cd melonDS
+   mkdir build
+   cd build
+   cmake .. -G 'MSYS Makefiles' -DBUILD_STATIC=ON -DQT5_STATIC_DIR=/mingw64/qt5-static
+   make -j$(nproc --all)
+   mkdir dist && cp melonDS.exe dist
+   ```
+If everything went well, melonDS should now be in the `dist` folder.
+
+### macOS:
+1. Install the [Homebrew Package Manager](https://brew.sh)
+2. Install dependencies: `brew install git pkg-config cmake sdl2 qt5 libslirp`
+3. Compile:
+   ```zsh
+   git clone https://github.com/Arisotura/melonDS.git
+   cd melonDS
+   mkdir build && cd build
+   cmake .. -DQt5_DIR=$(brew --prefix qt5)/lib/cmake/Qt5
+   make -j$(sysctl -n hw.ncpu)
+   mkdir dist && cp -r melonDS.app dist
+   ```
+If everything went well, melonDS.app should now be in the `dist` folder.
+
+   
 ## TODO LIST

- * better DSi emulation
- * better OpenGL rendering
- * netplay
+ * DSi emulation
 * the impossible quest of pixel-perfect 3D graphics
+ * improve libui and the emulator UI
 * support for rendering screens to separate windows
 * emulating some fancy addons
- * other non-core shit (debugger, graphics viewers, etc)
+ * other non-core shit (debugger, graphics viewers, cheat crapo, etc)

-### TODO LIST FOR LATER (low priority)
+### TODO LIST FOR LATER

+ * better wifi
+ * maybe emulate flashcarts or other fancy hardware
 * big-endian compatibility (Wii, etc)
 * LCD refresh time (used by some games for blending effects)
 * any feature you can eventually ask for that isn't outright stupid
@ -57,14 +123,10 @@ See [BUILD.md](./BUILD.md) for build instructions.
 * limittox for the icon
 * All of you comrades who have been testing melonDS, reporting issues, suggesting shit, etc

-## Licenses
-
+## License
 [![GNU GPLv3 Image](https://www.gnu.org/graphics/gplv3-127x51.png)](http://www.gnu.org/licenses/gpl-3.0.en.html)

 melonDS is free software: you can redistribute it and/or modify
 it under the terms of the GNU General Public License as published by
 the Free Software Foundation, either version 3 of the License, or
 (at your option) any later version.
-
-### External
-* Images used in the Input Config Dialog - see `src/frontend/qt_sdl/InputConfig/resources/LICENSE.md`
--- a/cmake/ConfigureVcpkg.cmake
+++ b/cmake/ConfigureVcpkg.cmake
@ -1,115 +0,0 @@
-include(FetchContent)
-
-set(_DEFAULT_VCPKG_ROOT "${CMAKE_SOURCE_DIR}/vcpkg")
-set(VCPKG_ROOT "${_DEFAULT_VCPKG_ROOT}" CACHE STRING "The path to the vcpkg repository")
-
-if (VCPKG_ROOT STREQUAL "${_DEFAULT_VCPKG_ROOT}")
-    if (APPLE) # this doesn't work on non-macOS
-        file(LOCK "${_DEFAULT_VCPKG_ROOT}" DIRECTORY GUARD FILE)
-    endif()
-    FetchContent_Declare(vcpkg
-        GIT_REPOSITORY "https://github.com/Microsoft/vcpkg.git"
-        GIT_TAG 2ad004460f5db4d3b66f62f5799ff66c265c4b5d
-        EXCLUDE_FROM_ALL
-        SOURCE_DIR "${CMAKE_SOURCE_DIR}/vcpkg")
-    FetchContent_MakeAvailable(vcpkg)
-endif()
-
-set(VCPKG_OVERLAY_TRIPLETS "${CMAKE_SOURCE_DIR}/cmake/overlay-triplets")
-
-option(USE_RECOMMENDED_TRIPLETS "Use the recommended triplets that are used for official builds" ON)
-
-# Duplicated here because it needs to be set before project()
-option(USE_QT6 "Use Qt 6 instead of Qt 5" ON)
-
-# Since the Linux build pulls in glib anyway, we can just use upstream libslirp
-if (UNIX AND NOT APPLE)
-    option(USE_SYSTEM_LIBSLIRP "Use system libslirp instead of the bundled version" ON)
-endif()
-
-if (NOT USE_QT6)
-    list(APPEND VCPKG_MANIFEST_FEATURES qt5)
-    set(VCPKG_MANIFEST_NO_DEFAULT_FEATURES ON)
-endif()
-
-if (CMAKE_OSX_ARCHITECTURES MATCHES ";")
-    message(FATAL_ERROR "macOS universal builds are not supported. Build them individually and combine afterwards instead.")
-endif()
-
-if (USE_RECOMMENDED_TRIPLETS)
-    execute_process(
-        COMMAND uname -m
-        OUTPUT_VARIABLE _HOST_PROCESSOR
-        OUTPUT_STRIP_TRAILING_WHITESPACE)
-
-    set(_CAN_TARGET_AS_HOST OFF)
-
-    if (APPLE)
-        if (NOT CMAKE_OSX_ARCHITECTURES)
-            if (_HOST_PROCESSOR STREQUAL arm64)
-                set(CMAKE_OSX_ARCHITECTURES arm64)
-            else()
-                set(CMAKE_OSX_ARCHITECTURES x86_64)
-            endif()
-        endif()
-
-        if (CMAKE_OSX_ARCHITECTURES STREQUAL arm64)
-            set(_WANTED_TRIPLET arm64-osx-11-release)
-            set(CMAKE_OSX_DEPLOYMENT_TARGET 11.0)
-        else()
-            set(_WANTED_TRIPLET x64-osx-1015-release)
-            set(CMAKE_OSX_DEPLOYMENT_TARGET 10.15)
-        endif()
-    elseif(WIN32)
-        # TODO Windows arm64 if possible
-        set(_CAN_TARGET_AS_HOST ON)
-        set(_WANTED_TRIPLET x64-mingw-static-release)
-    elseif(CMAKE_HOST_SYSTEM_NAME STREQUAL Linux)
-        # Can't really detect cross compiling here.
-        set(_CAN_TARGET_AS_HOST ON)
-        if (_HOST_PROCESSOR STREQUAL x86_64)
-            set(_WANTED_TRIPLET x64-linux-release)
-        elseif(_HOST_PROCESSOR STREQUAL "aarch64")
-            set(_WANTED_TRIPLET arm64-linux-release)
-        endif()
-    endif()
-
-    # Don't override it if the user set something else
-    if (NOT VCPKG_TARGET_TRIPLET)
-        set(VCPKG_TARGET_TRIPLET "${_WANTED_TRIPLET}")
-    else()
-        set(_WANTED_TRIPLET "${VCPKG_TARGET_TRIPLET}")
-    endif()
-
-    if (APPLE)
-        if (_HOST_PROCESSOR MATCHES arm64)
-            if (_WANTED_TRIPLET MATCHES "^arm64-osx-")
-                set(_CAN_TARGET_AS_HOST ON)
-            elseif (_WANTED_TRIPLET STREQUAL "x64-osx-1015-release")
-                # Use the default triplet for when building for arm64
-                # because we're probably making a universal build
-                set(VCPKG_HOST_TRIPLET arm64-osx-11-release)
-            endif()
-        else()
-            if (_WANTED_TRIPLET MATCHES "^x64-osx-")
-                set(_CAN_TARGET_AS_HOST ON)
-            elseif (_WANTED_TRIPLET STREQUAL "arm64-osx-11-release")
-                set(VCPKG_HOST_TRIPLET x64-osx-1015-release)
-            endif()
-        endif()
-    endif()
-
-    # If host and target triplet differ, vcpkg seems to always assume that the host can't run the target's binaries.
-    # In cases like cross compiling from ARM -> Intel macOS, or target being an older version of the host OS, we *can* do that so the packages built targeting the host are redundant.
-    if (_CAN_TARGET_AS_HOST AND NOT VCPKG_HOST_TRIPLET)
-        option(VCPKG_TARGET_AS_HOST "Use the target as host triplet to speed up builds" ON)
-    else()
-        option(VCPKG_TARGET_AS_HOST "Use the target as host triplet to speed up builds" OFF)
-    endif()
-
-    if (VCPKG_TARGET_AS_HOST)
-        set(VCPKG_HOST_TRIPLET "${VCPKG_TARGET_TRIPLET}" CACHE STRING "Host triplet to use for vcpkg")
-    endif()
-endif()
-
-set(CMAKE_TOOLCHAIN_FILE "${VCPKG_ROOT}/scripts/buildsystems/vcpkg.cmake")
--- a/cmake/DefaultBuildFlags.cmake
+++ b/cmake/DefaultBuildFlags.cmake
@ -1,9 +0,0 @@
-if (CMAKE_C_COMPILER_ID STREQUAL GNU)
-	set(CMAKE_C_FLAGS_DEBUG_INIT "-g -Og")
-endif()
-if (CMAKE_CXX_COMPILER_ID STREQUAL GNU)
-	set(CMAKE_CXX_FLAGS_DEBUG_INIT "-g -Og")
-endif()
-
-string(REPLACE "-O2" "-O3" CMAKE_C_FLAGS_RELEASE_INIT "${CMAKE_C_FLAGS_RELEASE_INIT}")
-string(REPLACE "-O2" "-O3" CMAKE_CXX_FLAGS_RELEASE_INIT "${CMAKE_CXX_FLAGS_RELEASE_INIT}")
--- a/cmake/ECMFindModuleHelpers.cmake
+++ b/cmake/ECMFindModuleHelpers.cmake
@ -1,297 +0,0 @@
-#.rst:
-# ECMFindModuleHelpers
-# --------------------
-#
-# Helper macros for find modules: ecm_find_package_version_check(),
-# ecm_find_package_parse_components() and
-# ecm_find_package_handle_library_components().
-#
-# ::
-#
-#   ecm_find_package_version_check(<name>)
-#
-# Prints warnings if the CMake version or the project's required CMake version
-# is older than that required by extra-cmake-modules.
-#
-# ::
-#
-#   ecm_find_package_parse_components(<name>
-#       RESULT_VAR <variable>
-#       KNOWN_COMPONENTS <component1> [<component2> [...]]
-#       [SKIP_DEPENDENCY_HANDLING])
-#
-# This macro will populate <variable> with a list of components found in
-# <name>_FIND_COMPONENTS, after checking that all those components are in the
-# list of KNOWN_COMPONENTS; if there are any unknown components, it will print
-# an error or warning (depending on the value of <name>_FIND_REQUIRED) and call
-# return().
-#
-# The order of components in <variable> is guaranteed to match the order they
-# are listed in the KNOWN_COMPONENTS argument.
-#
-# If SKIP_DEPENDENCY_HANDLING is not set, for each component the variable
-# <name>_<component>_component_deps will be checked for dependent components.
-# If <component> is listed in <name>_FIND_COMPONENTS, then all its (transitive)
-# dependencies will also be added to <variable>.
-#
-# ::
-#
-#   ecm_find_package_handle_library_components(<name>
-#       COMPONENTS <component> [<component> [...]]
-#       [SKIP_DEPENDENCY_HANDLING])
-#       [SKIP_PKG_CONFIG])
-#
-# Creates an imported library target for each component.  The operation of this
-# macro depends on the presence of a number of CMake variables.
-#
-# The <name>_<component>_lib variable should contain the name of this library,
-# and <name>_<component>_header variable should contain the name of a header
-# file associated with it (whatever relative path is normally passed to
-# '#include'). <name>_<component>_header_subdir variable can be used to specify
-# which subdirectory of the include path the headers will be found in.
-# ecm_find_package_components() will then search for the library
-# and include directory (creating appropriate cache variables) and create an
-# imported library target named <name>::<component>.
-#
-# Additional variables can be used to provide additional information:
-#
-# If SKIP_PKG_CONFIG, the <name>_<component>_pkg_config variable is set, and
-# pkg-config is found, the pkg-config module given by
-# <name>_<component>_pkg_config will be searched for and used to help locate the
-# library and header file.  It will also be used to set
-# <name>_<component>_VERSION.
-#
-# Note that if version information is found via pkg-config,
-# <name>_<component>_FIND_VERSION can be set to require a particular version
-# for each component.
-#
-# If SKIP_DEPENDENCY_HANDLING is not set, the INTERFACE_LINK_LIBRARIES property
-# of the imported target for <component> will be set to contain the imported
-# targets for the components listed in <name>_<component>_component_deps.
-# <component>_FOUND will also be set to false if any of the compoments in
-# <name>_<component>_component_deps are not found.  This requires the components
-# in <name>_<component>_component_deps to be listed before <component> in the
-# COMPONENTS argument.
-#
-# The following variables will be set:
-#
-# ``<name>_TARGETS``
-#   the imported targets
-# ``<name>_LIBRARIES``
-#   the found libraries
-# ``<name>_INCLUDE_DIRS``
-#   the combined required include directories for the components
-# ``<name>_DEFINITIONS``
-#   the "other" CFLAGS provided by pkg-config, if any
-# ``<name>_VERSION``
-#   the value of ``<name>_<component>_VERSION`` for the first component that
-#   has this variable set (note that components are searched for in the order
-#   they are passed to the macro), although if it is already set, it will not
-#   be altered
-#
-# Note that these variables are never cleared, so if
-# ecm_find_package_handle_library_components() is called multiple times with
-# different components (typically because of multiple find_package() calls) then
-# ``<name>_TARGETS``, for example, will contain all the targets found in any
-# call (although no duplicates).
-#
-# Since pre-1.0.0.
-
-#=============================================================================
-# Copyright 2014 Alex Merry <alex.merry@kde.org>
-#
-# Redistribution and use in source and binary forms, with or without
-# modification, are permitted provided that the following conditions
-# are met:
-#
-# 1. Redistributions of source code must retain the copyright
-#    notice, this list of conditions and the following disclaimer.
-# 2. Redistributions in binary form must reproduce the copyright
-#    notice, this list of conditions and the following disclaimer in the
-#    documentation and/or other materials provided with the distribution.
-# 3. The name of the author may not be used to endorse or promote products
-#    derived from this software without specific prior written permission.
-#
-# THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
-# IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
-# OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
-# IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
-# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
-# NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
-# THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-
-include(CMakeParseArguments)
-
-macro(ecm_find_package_version_check module_name)
-    if(CMAKE_VERSION VERSION_LESS 2.8.12)
-        message(FATAL_ERROR "CMake 2.8.12 is required by Find${module_name}.cmake")
-    endif()
-    if(CMAKE_MINIMUM_REQUIRED_VERSION VERSION_LESS 2.8.12)
-        message(AUTHOR_WARNING "Your project should require at least CMake 2.8.12 to use Find${module_name}.cmake")
-    endif()
-endmacro()
-
-macro(ecm_find_package_parse_components module_name)
-    set(ecm_fppc_options SKIP_DEPENDENCY_HANDLING)
-    set(ecm_fppc_oneValueArgs RESULT_VAR)
-    set(ecm_fppc_multiValueArgs KNOWN_COMPONENTS DEFAULT_COMPONENTS)
-    cmake_parse_arguments(ECM_FPPC "${ecm_fppc_options}" "${ecm_fppc_oneValueArgs}" "${ecm_fppc_multiValueArgs}" ${ARGN})
-
-    if(ECM_FPPC_UNPARSED_ARGUMENTS)
-        message(FATAL_ERROR "Unexpected arguments to ecm_find_package_parse_components: ${ECM_FPPC_UNPARSED_ARGUMENTS}")
-    endif()
-    if(NOT ECM_FPPC_RESULT_VAR)
-        message(FATAL_ERROR "Missing RESULT_VAR argument to ecm_find_package_parse_components")
-    endif()
-    if(NOT ECM_FPPC_KNOWN_COMPONENTS)
-        message(FATAL_ERROR "Missing KNOWN_COMPONENTS argument to ecm_find_package_parse_components")
-    endif()
-    if(NOT ECM_FPPC_DEFAULT_COMPONENTS)
-        set(ECM_FPPC_DEFAULT_COMPONENTS ${ECM_FPPC_KNOWN_COMPONENTS})
-    endif()
-
-    if(${module_name}_FIND_COMPONENTS)
-        set(ecm_fppc_requestedComps ${${module_name}_FIND_COMPONENTS})
-
-        if(NOT ECM_FPPC_SKIP_DEPENDENCY_HANDLING)
-            # Make sure deps are included
-            foreach(ecm_fppc_comp ${ecm_fppc_requestedComps})
-                foreach(ecm_fppc_dep_comp ${${module_name}_${ecm_fppc_comp}_component_deps})
-                    list(FIND ecm_fppc_requestedComps "${ecm_fppc_dep_comp}" ecm_fppc_index)
-                    if("${ecm_fppc_index}" STREQUAL "-1")
-                        if(NOT ${module_name}_FIND_QUIETLY)
-                            message(STATUS "${module_name}: ${ecm_fppc_comp} requires ${${module_name}_${ecm_fppc_comp}_component_deps}")
-                        endif()
-                        list(APPEND ecm_fppc_requestedComps "${ecm_fppc_dep_comp}")
-                    endif()
-                endforeach()
-            endforeach()
-        else()
-            message(STATUS "Skipping dependency handling for ${module_name}")
-        endif()
-        list(REMOVE_DUPLICATES ecm_fppc_requestedComps)
-
-        # This makes sure components are listed in the same order as
-        # KNOWN_COMPONENTS (potentially important for inter-dependencies)
-        set(${ECM_FPPC_RESULT_VAR})
-        foreach(ecm_fppc_comp ${ECM_FPPC_KNOWN_COMPONENTS})
-            list(FIND ecm_fppc_requestedComps "${ecm_fppc_comp}" ecm_fppc_index)
-            if(NOT "${ecm_fppc_index}" STREQUAL "-1")
-                list(APPEND ${ECM_FPPC_RESULT_VAR} "${ecm_fppc_comp}")
-                list(REMOVE_AT ecm_fppc_requestedComps ${ecm_fppc_index})
-            endif()
-        endforeach()
-        # if there are any left, they are unknown components
-        if(ecm_fppc_requestedComps)
-            set(ecm_fppc_msgType STATUS)
-            if(${module_name}_FIND_REQUIRED)
-                set(ecm_fppc_msgType FATAL_ERROR)
-            endif()
-            if(NOT ${module_name}_FIND_QUIETLY)
-                message(${ecm_fppc_msgType} "${module_name}: requested unknown components ${ecm_fppc_requestedComps}")
-            endif()
-            return()
-        endif()
-    else()
-        set(${ECM_FPPC_RESULT_VAR} ${ECM_FPPC_DEFAULT_COMPONENTS})
-    endif()
-endmacro()
-
-macro(ecm_find_package_handle_library_components module_name)
-    set(ecm_fpwc_options SKIP_PKG_CONFIG SKIP_DEPENDENCY_HANDLING)
-    set(ecm_fpwc_oneValueArgs)
-    set(ecm_fpwc_multiValueArgs COMPONENTS)
-    cmake_parse_arguments(ECM_FPWC "${ecm_fpwc_options}" "${ecm_fpwc_oneValueArgs}" "${ecm_fpwc_multiValueArgs}" ${ARGN})
-
-    if(ECM_FPWC_UNPARSED_ARGUMENTS)
-        message(FATAL_ERROR "Unexpected arguments to ecm_find_package_handle_components: ${ECM_FPWC_UNPARSED_ARGUMENTS}")
-    endif()
-    if(NOT ECM_FPWC_COMPONENTS)
-        message(FATAL_ERROR "Missing COMPONENTS argument to ecm_find_package_handle_components")
-    endif()
-
-    include(FindPackageHandleStandardArgs)
-    find_package(PkgConfig)
-    foreach(ecm_fpwc_comp ${ECM_FPWC_COMPONENTS})
-        set(ecm_fpwc_dep_vars)
-        set(ecm_fpwc_dep_targets)
-        if(NOT SKIP_DEPENDENCY_HANDLING)
-            foreach(ecm_fpwc_dep ${${module_name}_${ecm_fpwc_comp}_component_deps})
-                list(APPEND ecm_fpwc_dep_vars "${module_name}_${ecm_fpwc_dep}_FOUND")
-                list(APPEND ecm_fpwc_dep_targets "${module_name}::${ecm_fpwc_dep}")
-            endforeach()
-        endif()
-
-        if(NOT ECM_FPWC_SKIP_PKG_CONFIG AND ${module_name}_${ecm_fpwc_comp}_pkg_config)
-            pkg_check_modules(PKG_${module_name}_${ecm_fpwc_comp} QUIET
-                              ${${module_name}_${ecm_fpwc_comp}_pkg_config})
-        endif()
-
-        find_path(${module_name}_${ecm_fpwc_comp}_INCLUDE_DIR
-            NAMES ${${module_name}_${ecm_fpwc_comp}_header}
-            HINTS ${PKG_${module_name}_${ecm_fpwc_comp}_INCLUDE_DIRS}
-            PATH_SUFFIXES ${${module_name}_${ecm_fpwc_comp}_header_subdir}
-        )
-        find_library(${module_name}_${ecm_fpwc_comp}_LIBRARY
-            NAMES ${${module_name}_${ecm_fpwc_comp}_lib}
-            HINTS ${PKG_${module_name}_${ecm_fpwc_comp}_LIBRARY_DIRS}
-        )
-
-        set(${module_name}_${ecm_fpwc_comp}_VERSION "${PKG_${module_name}_${ecm_fpwc_comp}_VERSION}")
-        if(NOT ${module_name}_VERSION)
-            set(${module_name}_VERSION ${${module_name}_${ecm_fpwc_comp}_VERSION})
-        endif()
-
-        find_package_handle_standard_args(${module_name}_${ecm_fpwc_comp}
-            FOUND_VAR
-                ${module_name}_${ecm_fpwc_comp}_FOUND
-            REQUIRED_VARS
-                ${module_name}_${ecm_fpwc_comp}_LIBRARY
-                ${module_name}_${ecm_fpwc_comp}_INCLUDE_DIR
-                ${ecm_fpwc_dep_vars}
-            VERSION_VAR
-                ${module_name}_${ecm_fpwc_comp}_VERSION
-            )
-
-        mark_as_advanced(
-            ${module_name}_${ecm_fpwc_comp}_LIBRARY
-            ${module_name}_${ecm_fpwc_comp}_INCLUDE_DIR
-        )
-
-        if(${module_name}_${ecm_fpwc_comp}_FOUND)
-            list(APPEND ${module_name}_LIBRARIES
-                        "${${module_name}_${ecm_fpwc_comp}_LIBRARY}")
-            list(APPEND ${module_name}_INCLUDE_DIRS
-                        "${${module_name}_${ecm_fpwc_comp}_INCLUDE_DIR}")
-            set(${module_name}_DEFINITIONS
-                    ${${module_name}_DEFINITIONS}
-                    ${PKG_${module_name}_${ecm_fpwc_comp}_DEFINITIONS})
-            if(NOT TARGET ${module_name}::${ecm_fpwc_comp})
-                add_library(${module_name}::${ecm_fpwc_comp} UNKNOWN IMPORTED)
-                set_target_properties(${module_name}::${ecm_fpwc_comp} PROPERTIES
-                    IMPORTED_LOCATION "${${module_name}_${ecm_fpwc_comp}_LIBRARY}"
-                    INTERFACE_COMPILE_OPTIONS "${PKG_${module_name}_${ecm_fpwc_comp}_DEFINITIONS}"
-                    INTERFACE_INCLUDE_DIRECTORIES "${${module_name}_${ecm_fpwc_comp}_INCLUDE_DIR}"
-                    INTERFACE_LINK_LIBRARIES "${ecm_fpwc_dep_targets}"
-                )
-            endif()
-            list(APPEND ${module_name}_TARGETS
-                        "${module_name}::${ecm_fpwc_comp}")
-        endif()
-    endforeach()
-    if(${module_name}_LIBRARIES)
-        list(REMOVE_DUPLICATES ${module_name}_LIBRARIES)
-    endif()
-    if(${module_name}_INCLUDE_DIRS)
-        list(REMOVE_DUPLICATES ${module_name}_INCLUDE_DIRS)
-    endif()
-    if(${module_name}_DEFINITIONS)
-        list(REMOVE_DUPLICATES ${module_name}_DEFINITIONS)
-    endif()
-    if(${module_name}_TARGETS)
-        list(REMOVE_DUPLICATES ${module_name}_TARGETS)
-    endif()
-endmacro()
--- a/cmake/ECMFindModuleHelpersStub.cmake
+++ b/cmake/ECMFindModuleHelpersStub.cmake
@ -1 +0,0 @@
-include(${CMAKE_CURRENT_LIST_DIR}/ECMFindModuleHelpers.cmake)
--- a/cmake/FindEGL.cmake
+++ b/cmake/FindEGL.cmake
@ -1,172 +0,0 @@
-#.rst:
-# FindEGL
-# -------
-#
-# Try to find EGL.
-#
-# This will define the following variables:
-#
-# ``EGL_FOUND``
-#     True if (the requested version of) EGL is available
-# ``EGL_VERSION``
-#     The version of EGL; note that this is the API version defined in the
-#     headers, rather than the version of the implementation (eg: Mesa)
-# ``EGL_LIBRARIES``
-#     This can be passed to target_link_libraries() instead of the ``EGL::EGL``
-#     target
-# ``EGL_INCLUDE_DIRS``
-#     This should be passed to target_include_directories() if the target is not
-#     used for linking
-# ``EGL_DEFINITIONS``
-#     This should be passed to target_compile_options() if the target is not
-#     used for linking
-#
-# If ``EGL_FOUND`` is TRUE, it will also define the following imported target:
-#
-# ``EGL::EGL``
-#     The EGL library
-#
-# In general we recommend using the imported target, as it is easier to use.
-# Bear in mind, however, that if the target is in the link interface of an
-# exported library, it must be made available by the package config file.
-#
-# Since pre-1.0.0.
-
-#=============================================================================
-# Copyright 2014 Alex Merry <alex.merry@kde.org>
-# Copyright 2014 Martin Gräßlin <mgraesslin@kde.org>
-#
-# Redistribution and use in source and binary forms, with or without
-# modification, are permitted provided that the following conditions
-# are met:
-#
-# 1. Redistributions of source code must retain the copyright
-#    notice, this list of conditions and the following disclaimer.
-# 2. Redistributions in binary form must reproduce the copyright
-#    notice, this list of conditions and the following disclaimer in the
-#    documentation and/or other materials provided with the distribution.
-# 3. The name of the author may not be used to endorse or promote products
-#    derived from this software without specific prior written permission.
-#
-# THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
-# IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
-# OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
-# IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
-# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
-# NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
-# THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-#=============================================================================
-
-include(${CMAKE_CURRENT_LIST_DIR}/ECMFindModuleHelpersStub.cmake)
-include(CheckCXXSourceCompiles)
-include(CMakePushCheckState)
-
-ecm_find_package_version_check(EGL)
-
-# Use pkg-config to get the directories and then use these values
-# in the FIND_PATH() and FIND_LIBRARY() calls
-find_package(PkgConfig)
-pkg_check_modules(PKG_EGL QUIET egl)
-
-set(EGL_DEFINITIONS ${PKG_EGL_CFLAGS_OTHER})
-
-find_path(EGL_INCLUDE_DIR
-    NAMES
-        EGL/egl.h
-    HINTS
-        ${PKG_EGL_INCLUDE_DIRS}
-)
-find_library(EGL_LIBRARY
-    NAMES
-        EGL
-    HINTS
-        ${PKG_EGL_LIBRARY_DIRS}
-)
-
-# NB: We do *not* use the version information from pkg-config, as that
-#     is the implementation version (eg: the Mesa version)
-if(EGL_INCLUDE_DIR)
-    # egl.h has defines of the form EGL_VERSION_x_y for each supported
-    # version; so the header for EGL 1.1 will define EGL_VERSION_1_0 and
-    # EGL_VERSION_1_1.  Finding the highest supported version involves
-    # finding all these defines and selecting the highest numbered.
-    file(READ "${EGL_INCLUDE_DIR}/EGL/egl.h" _EGL_header_contents)
-    string(REGEX MATCHALL
-        "[ \t]EGL_VERSION_[0-9_]+"
-        _EGL_version_lines
-        "${_EGL_header_contents}"
-    )
-    unset(_EGL_header_contents)
-    foreach(_EGL_version_line ${_EGL_version_lines})
-        string(REGEX REPLACE
-            "[ \t]EGL_VERSION_([0-9_]+)"
-            "\\1"
-            _version_candidate
-            "${_EGL_version_line}"
-        )
-        string(REPLACE "_" "." _version_candidate "${_version_candidate}")
-        if(NOT DEFINED EGL_VERSION OR EGL_VERSION VERSION_LESS _version_candidate)
-            set(EGL_VERSION "${_version_candidate}")
-        endif()
-    endforeach()
-    unset(_EGL_version_lines)
-endif()
-
-cmake_push_check_state(RESET)
-list(APPEND CMAKE_REQUIRED_LIBRARIES "${EGL_LIBRARY}")
-list(APPEND CMAKE_REQUIRED_INCLUDES "${EGL_INCLUDE_DIR}")
-
-check_cxx_source_compiles("
-#include <EGL/egl.h>
-
-int main(int argc, char *argv[]) {
-    EGLint x = 0; EGLDisplay dpy = 0; EGLContext ctx = 0;
-    eglDestroyContext(dpy, ctx);
-}" HAVE_EGL)
-
-cmake_pop_check_state()
-
-set(required_vars EGL_INCLUDE_DIR HAVE_EGL)
-if(NOT EMSCRIPTEN)
-    list(APPEND required_vars EGL_LIBRARY)
-endif()
-
-include(FindPackageHandleStandardArgs)
-find_package_handle_standard_args(EGL
-    FOUND_VAR
-        EGL_FOUND
-    REQUIRED_VARS
-        ${required_vars}
-    VERSION_VAR
-        EGL_VERSION
-)
-
-if(EGL_FOUND AND NOT TARGET EGL::EGL)
-    if (EMSCRIPTEN)
-        add_library(EGL::EGL INTERFACE IMPORTED)
-        # Nothing further to be done, system include paths have headers and linkage is implicit.
-    else()
-        add_library(EGL::EGL UNKNOWN IMPORTED)
-        set_target_properties(EGL::EGL PROPERTIES
-            IMPORTED_LOCATION "${EGL_LIBRARY}"
-            INTERFACE_COMPILE_OPTIONS "${EGL_DEFINITIONS}"
-            INTERFACE_INCLUDE_DIRECTORIES "${EGL_INCLUDE_DIR}"
-        )
-    endif()
-endif()
-
-mark_as_advanced(EGL_LIBRARY EGL_INCLUDE_DIR HAVE_EGL)
-
-# compatibility variables
-set(EGL_LIBRARIES ${EGL_LIBRARY})
-set(EGL_INCLUDE_DIRS ${EGL_INCLUDE_DIR})
-set(EGL_VERSION_STRING ${EGL_VERSION})
-
-include(FeatureSummary)
-set_package_properties(EGL PROPERTIES
-    URL "https://www.khronos.org/egl/"
-    DESCRIPTION "A platform-agnostic mechanism for creating rendering surfaces for use with other graphics libraries, such as OpenGL|ES and OpenVG."
-)
--- a/cmake/FindENet.cmake
+++ b/cmake/FindENet.cmake
@ -1,48 +0,0 @@
-# - Try to find enet
-# Once done this will define
-#
-#  ENET_FOUND - system has enet
-#  ENET_INCLUDE_DIRS - the enet include directory
-#  ENET_LIBRARIES - the libraries needed to use enet
-#
-# $ENETDIR is an environment variable used for finding enet.
-#
-#  Borrowed from The Mana World
-#  http://themanaworld.org/
-#
-# Several changes and additions by Fabian 'x3n' Landau
-# Lots of simplifications by Adrian Friedli
-#                 > www.orxonox.net <
-
-FIND_PATH(ENET_INCLUDE_DIRS enet/enet.h
-    PATHS
-    $ENV{ENETDIR}
-    /usr/local
-    /usr
-    PATH_SUFFIXES include
-)
-
-FIND_LIBRARY(ENET_LIBRARY
-    NAMES enet
-    PATHS
-    $ENV{ENETDIR}
-    /usr/local
-    /usr
-    PATH_SUFFIXES lib
-)
-
-# handle the QUIETLY and REQUIRED arguments and set ENET_FOUND to TRUE if
-# all listed variables are TRUE
-INCLUDE(FindPackageHandleStandardArgs)
-FIND_PACKAGE_HANDLE_STANDARD_ARGS(ENet DEFAULT_MSG ENET_LIBRARY ENET_INCLUDE_DIRS)
-
-IF (ENET_FOUND)
-    IF(WIN32)
-        SET(WINDOWS_ENET_DEPENDENCIES "ws2_32;winmm")
-        SET(ENET_LIBRARIES ${ENET_LIBRARY} ${WINDOWS_ENET_DEPENDENCIES})
-    ELSE(WIN32)
-        SET(ENET_LIBRARIES ${ENET_LIBRARY})
-    ENDIF(WIN32)
-ENDIF (ENET_FOUND)
-
-MARK_AS_ADVANCED(ENET_LIBRARY ENET_LIBRARIES ENET_INCLUDE_DIRS)
--- a/cmake/FindVTune.cmake
+++ b/cmake/FindVTune.cmake
@ -1,10 +0,0 @@
-
-find_path(VTUNE_PATH "")
-
-set(VTUNE_INCLUDE_DIR "${VTUNE_PATH}/include")
-
-if (WIN32)
-    set(VTUNE_LIBRARY "${VTUNE_PATH}/lib64/jitprofiling.lib")
-else()
-    set(VTUNE_LIBRARY "${VTUNE_PATH}/lib64/jitprofiling.a")
-endif()
--- a/cmake/FixInterfaceIncludes.cmake
+++ b/cmake/FixInterfaceIncludes.cmake
@ -1,35 +0,0 @@
-# The entire codebase quite reasonably does things like #include <SDL2/SDL.h> or <epoxy/gl.h>
-# CMake apparently doesn't think you should be doing this, so just includes $PREFIX/include/packagename for a given
-# package as include directories when using `target_link_libraries` with an imported target, this hacky function fixes
-# that up so includes can keep working as they always did but we can still use fancy imported targets.
-# This is stupid.
-
-function(fix_interface_includes)
-    foreach (target ${ARGN})
-        set(NEW_DIRS)
-        get_target_property(DIRS "${target}" INTERFACE_INCLUDE_DIRECTORIES)
-
-        if (NOT DIRS)
-            continue()
-        endif()
-
-        foreach (DIR ${DIRS})
-            get_filename_component(PARENT_DIR "${DIR}" DIRECTORY)
-
-            if (PARENT_DIR MATCHES "include$")
-                list(APPEND NEW_DIRS "${PARENT_DIR}")
-            endif()
-
-            # HACK
-            # The libarchive pkg-config file in MSYS2 seems to include a UNIX-style path for its
-            # include directory and CMake doesn't like that.
-            if (WIN32 AND MINGW AND target STREQUAL PkgConfig::LibArchive)
-                list(FILTER DIRS EXCLUDE REGEX "^/[^.]+64/.*")
-            endif()
-        endforeach()
-
-        list(APPEND DIRS ${NEW_DIRS})
-        set_target_properties("${target}" PROPERTIES INTERFACE_INCLUDE_DIRECTORIES "${DIRS}")
-    endforeach()
-endfunction()
-
--- a/cmake/Sanitizers.cmake
+++ b/cmake/Sanitizers.cmake
@ -1,8 +0,0 @@
-set(SANITIZE "" CACHE STRING "Sanitizers to enable.")
-
-string(REGEX MATCHALL "[^,]+" ENABLED_SANITIZERS "${SANITIZE}")
-
-foreach(SANITIZER ${ENABLED_SANITIZERS})
-    add_compile_options("-fsanitize=${SANITIZER}")
-    add_link_options("-fsanitize=${SANITIZER}")
-endforeach()
--- a/cmake/SetupCCache.cmake
+++ b/cmake/SetupCCache.cmake
@ -1,19 +0,0 @@
-include(FindPackageMessage)
-
-find_program(CCACHE "ccache")
-
-cmake_dependent_option(USE_CCACHE "Use CCache to speed up repeated builds." ON CCACHE OFF)
-
-if (NOT CCACHE OR NOT USE_CCACHE)
-    return()
-endif()
-
-# Fedora, and probably also Red Hat-based distros in general, use CCache by default if it's installed on the system.
-# We'll try to detect this here, and exit if that's the case.
-# Trying to launch ccache with ccache as we'd otherwise do seems to cause build issues.
-if (CMAKE_C_COMPILER MATCHES "ccache" OR CMAKE_CXX_COMPILER MATCHES "ccache")
-    return()
-endif()
-
-find_package_message(CCache "Using CCache to speed up compilation" "${USE_CCACHE}")
-set_property(GLOBAL PROPERTY RULE_LAUNCH_COMPILE "${CCACHE}")
--- a/cmake/Toolchain-Homebrew-LLVM.cmake
+++ b/cmake/Toolchain-Homebrew-LLVM.cmake
@ -1,12 +0,0 @@
-# Toolchain file for building with Homebrew's LLVM on macOS
-# This is useful on 10.14 where std::filesystem is not supported.
-
-set(CMAKE_C_COMPILER /usr/local/opt/llvm/bin/clang)
-set(CMAKE_CXX_COMPILER /usr/local/opt/llvm/bin/clang++)
-
-add_link_options(-L/usr/local/opt/llvm/lib)
-
-# LLVM in Homebrew is built with latest Xcode which has a newer linker than
-# what is bundled in the default install of Xcode Command Line Tools, so we
-# override it to prevent it passing flags not supported by the system's ld.
-add_link_options(-mlinker-version=450)
--- a/cmake/overlay-triplets/arm64-osx-11-release.cmake
+++ b/cmake/overlay-triplets/arm64-osx-11-release.cmake
@ -1,12 +0,0 @@
-set(VCPKG_TARGET_ARCHITECTURE arm64)
-set(VCPKG_CRT_LINKAGE dynamic)
-set(VCPKG_LIBRARY_LINKAGE static)
-
-set(VCPKG_CMAKE_SYSTEM_NAME Darwin)
-set(VCPKG_CMAKE_SYSTEM_VERSION 11.0)
-set(VCPKG_OSX_ARCHITECTURES arm64)
-set(VCPKG_BUILD_TYPE release)
-set(VCPKG_OSX_DEPLOYMENT_TARGET 11.0)
-
-set(VCPKG_C_FLAGS -mmacosx-version-min=11.0)
-set(VCPKG_CXX_FLAGS -mmacosx-version-min=11.0)
--- a/cmake/overlay-triplets/x64-mingw-static-release.cmake
+++ b/cmake/overlay-triplets/x64-mingw-static-release.cmake
@ -1,7 +0,0 @@
-set(VCPKG_TARGET_ARCHITECTURE x64)
-set(VCPKG_CRT_LINKAGE dynamic)
-set(VCPKG_LIBRARY_LINKAGE static)
-set(VCPKG_ENV_PASSTHROUGH PATH)
-set(VCPKG_BUILD_TYPE release)
-
-set(VCPKG_CMAKE_SYSTEM_NAME MinGW)
--- a/cmake/overlay-triplets/x64-osx-1015-release.cmake
+++ b/cmake/overlay-triplets/x64-osx-1015-release.cmake
@ -1,12 +0,0 @@
-set(VCPKG_TARGET_ARCHITECTURE x64)
-set(VCPKG_CRT_LINKAGE dynamic)
-set(VCPKG_LIBRARY_LINKAGE static)
-
-set(VCPKG_CMAKE_SYSTEM_NAME Darwin)
-set(VCPKG_CMAKE_SYSTEM_VERSION 10.15)
-set(VCPKG_OSX_ARCHITECTURES x86_64)
-set(VCPKG_BUILD_TYPE release)
-set(VCPKG_OSX_DEPLOYMENT_TARGET 10.15)
-
-set(VCPKG_C_FLAGS -mmacosx-version-min=10.15)
-set(VCPKG_CXX_FLAGS -mmacosx-version-min=10.15)
--- a/flake.lock
+++ b/flake.lock
@ -1,61 +0,0 @@
-{
-  "nodes": {
-    "flake-utils": {
-      "inputs": {
-        "systems": "systems"
-      },
-      "locked": {
-        "lastModified": 1731533236,
-        "narHash": "sha256-l0KFg5HjrsfsO/JpG+r7fRrqm12kzFHyUHqHCVpMMbI=",
-        "owner": "numtide",
-        "repo": "flake-utils",
-        "rev": "11707dc2f618dd54ca8739b309ec4fc024de578b",
-        "type": "github"
-      },
-      "original": {
-        "owner": "numtide",
-        "repo": "flake-utils",
-        "type": "github"
-      }
-    },
-    "nixpkgs": {
-      "locked": {
-        "lastModified": 1739020877,
-        "narHash": "sha256-mIvECo/NNdJJ/bXjNqIh8yeoSjVLAuDuTUzAo7dzs8Y=",
-        "owner": "NixOS",
-        "repo": "nixpkgs",
-        "rev": "a79cfe0ebd24952b580b1cf08cd906354996d547",
-        "type": "github"
-      },
-      "original": {
-        "owner": "NixOS",
-        "ref": "nixos-unstable",
-        "repo": "nixpkgs",
-        "type": "github"
-      }
-    },
-    "root": {
-      "inputs": {
-        "flake-utils": "flake-utils",
-        "nixpkgs": "nixpkgs"
-      }
-    },
-    "systems": {
-      "locked": {
-        "lastModified": 1681028828,
-        "narHash": "sha256-Vy1rq5AaRuLzOxct8nz4T6wlgyUR7zLU309k9mBC768=",
-        "owner": "nix-systems",
-        "repo": "default",
-        "rev": "da67096a3b9bf56a91d16901293e51ba5b49a27e",
-        "type": "github"
-      },
-      "original": {
-        "owner": "nix-systems",
-        "repo": "default",
-        "type": "github"
-      }
-    }
-  },
-  "root": "root",
-  "version": 7
-}
--- a/flake.nix
+++ b/flake.nix
@ -1,109 +0,0 @@
-{
-  description = "Nintendo DS emulator";
-
-  inputs = {
-    nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
-    flake-utils.url = "github:numtide/flake-utils";
-  };
-
-  outputs = { self, nixpkgs, flake-utils }: flake-utils.lib.eachDefaultSystem (system:
-    let
-      pkgs = import nixpkgs { inherit system; };
-      inherit (pkgs.lib) cmakeBool optionals makeLibraryPath;
-      inherit (pkgs.stdenv) isLinux isDarwin;
-
-      revision = with self; if sourceInfo?dirtyRev
-        then sourceInfo.dirtyRev
-        else sourceInfo.rev;
-      shortRevision = with self; if sourceInfo?dirtyShortRev
-        then sourceInfo.dirtyShortRev
-        else sourceInfo.shortRev;
-
-      melonDS = pkgs.stdenv.mkDerivation {
-        pname = "melonDS";
-        version = "1.0-${shortRevision}";
-        src = ./.;
-
-        nativeBuildInputs = with pkgs; [
-          cmake
-          ninja
-          pkg-config
-          qt6.wrapQtAppsHook
-        ];
-
-        buildInputs = (with pkgs; [
-          qt6.qtbase
-          qt6.qtmultimedia
-          SDL2
-          zstd
-          libarchive
-          libGL
-          libslirp
-          enet
-        ]) ++ optionals (!isDarwin) (with pkgs; [
-          kdePackages.extra-cmake-modules
-          qt6.qtwayland
-          wayland
-        ]);
-
-        cmakeFlags = [
-          (cmakeBool "USE_QT6" true)
-          (cmakeBool "USE_SYSTEM_LIBSLIRP" true)
-          (cmakeBool "MELONDS_EMBED_BUILD_INFO" true)
-        ];
-
-        env.MELONDS_GIT_HASH = revision;
-        env.MELONDS_GIT_BRANCH = "(unknown)";
-        env.MELONDS_BUILD_PROVIDER = "Nix";
-
-        qtWrapperArgs = optionals isLinux [
-          "--prefix LD_LIBRARY_PATH : ${makeLibraryPath [ pkgs.libpcap pkgs.wayland ]}"
-        ] ++ optionals isDarwin [
-          "--prefix DYLD_LIBRARY_PATH : ${makeLibraryPath [ pkgs.libpcap ]}"
-        ];
-
-        passthru = {
-          exePath = if isDarwin then
-            "/Applications/melonDS.app/Contents/MacOS/melonDS"
-            else "/bin/melonDS";
-        };
-      };
-    in {
-      packages.default = melonDS;
-      apps.default = flake-utils.lib.mkApp {
-        drv = self.packages.${system}.default;
-      };
-      devShells = {
-        default = pkgs.mkShell {
-          inputsFrom = [ self.packages.${system}.default ];
-          packages = with pkgs; [
-            qt6.qttools
-          ];
-        };
-
-        # Shell for building static melonDS release builds with vcpkg
-        # Use mkShellNoCC to ensure Nix's gcc/clang and stdlib isn't used
-        vcpkg = pkgs.mkShellNoCC {
-          packages = with pkgs; [
-            autoconf
-            autoconf-archive
-            automake
-            cmake
-            cups.dev # Needed by qtbase despite not enabling print support
-            git
-            iconv.dev
-            libtool
-            ninja
-            pkg-config
-            python3
-          ];
-
-          # Undo the SDK setup done by nixpkgs so we can use AppleClang
-          shellHook = ''
-            unset DEVELOPER_DIR SDKROOT MACOSX_DEPLOYMENT_TARGET
-         '';
-        };
-      };
-    }
-  );
-}
--- a/flatpak/net.kuribo64.melonDS.yml
+++ b/flatpak/net.kuribo64.melonDS.yml
@ -0,0 +1,29 @@
+---
+app-id: net.kuribo64.melonDS
+runtime: org.freedesktop.Platform
+runtime-version: '18.08'
+sdk: org.freedesktop.Sdk
+command: melonDS
+finish-args:
+  - "--share=ipc"
+  - "--socket=x11"
+  - "--socket=pulseaudio"
+  - "--share=network"
+  - "--device=all"
+  - "--filesystem=home"
+modules:
+  - name: libpcap
+    sources:
+      - type: archive
+        url: http://www.tcpdump.org/release/libpcap-1.9.0.tar.gz
+        sha256: 2edb88808e5913fdaa8e9c1fcaf272e19b2485338742b5074b9fe44d68f37019
+
+  - name: melonds
+    buildsystem: cmake-ninja
+    sources:
+      - type: git
+        url: https://github.com/Arisotura/melonDS.git
+        branch: master
+    post-install:
+      - "desktop-file-install --dir=/app/share/applications net.kuribo64.melonDS.desktop"
+      - "install -D icon/melon_256x256.png /app/share/icons/hicolor/256x256/apps/net.kuribo64.melonDS.png"
--- a/freebios/Makefile
+++ b/freebios/Makefile
@ -1,17 +0,0 @@
-TC_PREFIX  = /home/exophase/pandora-dev
-PREFIX     = $(TC_PREFIX)/arm-2011.03
-AS         = $(PREFIX)/bin/arm-none-linux-gnueabi-gcc
-OBJCOPY    = $(PREFIX)/bin/arm-none-linux-gnueabi-objcopy
-
-BIN_ARM7   = drastic_bios_arm7
-BIN_ARM9   = drastic_bios_arm9
-
-all:
-	$(AS) bios_common.S -DBIOS_ARM7 -march=armv4 -c -Wa,-asl=$(BIN_ARM7).list -o $(BIN_ARM7).o
-	$(AS) bios_common.S -DBIOS_ARM9 -march=armv5 -c -Wa,-asl=$(BIN_ARM9).list -o $(BIN_ARM9).o
-	$(OBJCOPY) -O binary $(BIN_ARM7).o $(BIN_ARM7).bin
-	$(OBJCOPY) -O binary $(BIN_ARM9).o $(BIN_ARM9).bin
-
-clean:
-	rm -f $(BIN_ARM7).bin $(BIN_ARM7).o $(BIN_ARM9).bin $(BIN_ARM9).o
-
--- a/freebios/bios_common.S
+++ b/freebios/bios_common.S
--- a/freebios/drastic_bios_arm7.bin
+++ b/freebios/drastic_bios_arm7.bin
--- a/freebios/drastic_bios_arm9.bin
+++ b/freebios/drastic_bios_arm9.bin
--- a/freebios/drastic_bios_readme.txt
+++ b/freebios/drastic_bios_readme.txt
@ -1,36 +0,0 @@
-Custom NDS ARM7/ARM9 BIOS replacement
-Copyright (c) 2013, Gilead Kutnick
-All rights reserved.
-
-Redistribution and use in source and binary forms, with or without
- modification, are permitted provided that the following conditions are met:
-
-1) Redistributions of source code must retain the above copyright notice,
-   this list of conditions and the following disclaimer.
-2) Redistributions in binary form must reproduce the above copyright notice,
-   this list of conditions and the following disclaimer in the documentation
-   and/or other materials provided with the distribution.
-
-THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
-AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
-IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
-ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
-LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
-CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
-SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
-INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
-CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
-ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
-POSSIBILITY OF SUCH DAMAGE.
-
-- Info --
-
-This archive contains source code and assembly for a custom BIOS replacement
-for the Nintendo DS system. This code is in no way affiliated with Nintendo
-and is not derived from Nintendo's BIOS implementation but has been implemented
-using publically available documentation.
-
-It can be assembled using the included Makefile along with a proper ARM gcc
-toolchain. Change the first four lines to point to the proper toolchain of your
-choice.
-
--- a/icon/melon_128x128.png
+++ b/icon/melon_128x128.png
--- a/icon/melon_16x16.png
+++ b/icon/melon_16x16.png
--- a/icon/melon_256x256.png
+++ b/icon/melon_256x256.png
--- a/icon/melon_32x32.png
+++ b/icon/melon_32x32.png
--- a/icon/melon_48x48.png
+++ b/icon/melon_48x48.png
--- a/icon/melon_64x64.png
+++ b/icon/melon_64x64.png
--- a/res/melon.ico
+++ b/res/melon.ico
--- a/melon.qrc
+++ b/melon.qrc
@ -0,0 +1,6 @@
+<!DOCTYPE RCC>
+<RCC version="1.0">
+	<qresource>
+		<file alias="melon-icon">icon/melon_32x32.png</file>
+	</qresource>
+</RCC>
--- a/res/melon.rc.in
+++ b/res/melon.rc.in
@ -2,12 +2,12 @@
 #define VFT_APP                      0x00000001L

 //this will set your .exe icon
-100 ICON MOVEABLE PURE LOADONCALL DISCARDABLE "${CMAKE_SOURCE_DIR}/res/melon.ico"
+100 ICON MOVEABLE PURE LOADONCALL DISCARDABLE "melon.ico"

 //include version information in .exe, modify these values to match your needs
 1 VERSIONINFO
-FILEVERSION ${MELON_RC_VERSION}
-PRODUCTVERSION ${MELON_RC_VERSION}
+FILEVERSION 0,9,1,0
+PRODUCTVERSION 0,9,1,0
 FILETYPE VFT_APP
 {
  BLOCK "StringFileInfo"
@ -15,14 +15,14 @@ FILETYPE VFT_APP
 		BLOCK "040904E4"
 		{
 			VALUE "CompanyName", "Melon Factory of Kuribo64"
-			VALUE "FileVersion", "${melonDS_VERSION}"
-			VALUE "FileDescription", "melonDS emulator"
+			VALUE "FileVersion", "0.9.1"
+			VALUE "FileDescription", "DS emulator, sorta. also 1st quality melon."
 			VALUE "InternalName", "SDnolem"
-			VALUE "LegalCopyright", "2016-2023 melonDS team"
+			VALUE "LegalCopyright", "2016-2020 Arisotura & co."
 			VALUE "LegalTrademarks", ""
-			VALUE "OriginalFilename", "melonDS.exe"
+			VALUE "OriginalFilename", "zafkflzdasd.exe"
 			VALUE "ProductName", "melonDS"
-			VALUE "ProductVersion", "${melonDS_VERSION}"
+			VALUE "ProductVersion", "0.9.1"
 		}
 	}
  BLOCK "VarFileInfo"
--- a/melonDS.icns
+++ b/melonDS.icns
--- a/melonDS.plist
+++ b/melonDS.plist
@ -0,0 +1,24 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+    <key>CFBundleExecutable</key>
+    <string>melonDS</string>
+    <key>CFBundleIconFile</key>
+    <string>melonDS.icns</string>
+    <key>CFBundleIdentifier</key>
+    <string>net.kuribo64.melonDS</string>
+    <key>CFBundleDevelopmentRegion</key>
+    <string>English</string>
+    <key>CFBundlePackageType</key>
+    <string>APPL</string>
+    <key>CFBundleVersion</key>
+    <string>0.9</string>
+    <key>CFBundleShortVersionString</key>
+    <string>0.9</string> 
+    <key>NSHumanReadableCopyright</key>
+    <string>Licensed under GPLv3</string>
+    <key>NSHighResolutionCapable</key>
+    <true/>
+</dict>
+</plist>
--- a/melon_grc.xml
+++ b/melon_grc.xml
@ -0,0 +1,11 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<gresources>
+  <gresource prefix="/org/kuriboland/melonDS">
+    <file preprocess="to-pixdata">icon/melon_16x16.png</file>
+    <file preprocess="to-pixdata">icon/melon_32x32.png</file>
+    <file preprocess="to-pixdata">icon/melon_48x48.png</file>
+    <file preprocess="to-pixdata">icon/melon_64x64.png</file>
+    <file preprocess="to-pixdata">icon/melon_128x128.png</file>
+    <file preprocess="to-pixdata">icon/melon_256x256.png</file>
+  </gresource>
+</gresources>
--- a/tools/msys-dist.sh
+++ b/tools/msys-dist.sh
--- a/res/net.kuribo64.melonDS.desktop
+++ b/res/net.kuribo64.melonDS.desktop
@ -2,7 +2,7 @@
 Name=melonDS
 GenericName=Nintendo DS Emulator
 Comment=A fast and accurate Nintendo DS emulator.
-Exec=melonDS %f
+Exec=melonDS
 Type=Application
 Categories=Game;Emulator;
 Terminal=false
--- a/res/icon/melon_128x128.png
+++ b/res/icon/melon_128x128.png
--- a/res/icon/melon_16x16.png
+++ b/res/icon/melon_16x16.png
--- a/res/icon/melon_192x192.png
+++ b/res/icon/melon_192x192.png
--- a/res/icon/melon_256x256.png
+++ b/res/icon/melon_256x256.png
--- a/res/icon/melon_32x32.png
+++ b/res/icon/melon_32x32.png
--- a/res/icon/melon_48x48.png
+++ b/res/icon/melon_48x48.png
--- a/res/icon/melon_64x64.png
+++ b/res/icon/melon_64x64.png
--- a/res/melon.icns
+++ b/res/melon.icns
--- a/res/melon.plist.in
+++ b/res/melon.plist.in
@ -1,101 +0,0 @@
-<?xml version="1.0" encoding="UTF-8"?>
-<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
-<plist version="1.0">
-<dict>
-    <key>CFBundleExecutable</key>
-    <string>melonDS</string>
-    <key>CFBundleIconFile</key>
-    <string>melon.icns</string>
-    <key>CFBundleIdentifier</key>
-    <string>net.kuribo64.melonDS</string>
-    <key>CFBundleDevelopmentRegion</key>
-    <string>English</string>
-    <key>CFBundlePackageType</key>
-    <string>APPL</string>
-    <key>CFBundleVersion</key>
-    <string>${melonDS_VERSION}</string>
-    <key>CFBundleShortVersionString</key>
-    <string>${melonDS_VERSION}</string> 
-    <key>LSApplicationCategoryType</key>
-    <string>public.app-category.games</string>
-    <key>NSHumanReadableCopyright</key>
-    <string>Licensed under GPLv3</string>
-    <key>NSPrincipalClass</key>
-    <string>NSApplication</string>
-    <key>NSHighResolutionCapable</key>
-    <true/>
-    <key>NSMicrophoneUsageDescription</key>
-    <string>We need microphone access so you can use the emulated DS microphone</string>
-    <key>NSCameraUsageDescription</key>
-    <string>Camera access is needed for emulation of the DSi's cameras</string>
-    <key>CFBundleDocumentTypes</key>
-    <array>
-        <dict>
-            <key>CFBundleTypeName</key>
-            <string>Nintendo DS ROM</string>
-            <key>CFBundleTypeExtensions</key>
-            <array>
-                <string>nds</string>
-                <string>srl</string>
-                <string>dsi</string>
-                <string>ids</string>
-                <string>nds.zst</string>
-                <string>srl.zst</string>
-                <string>dsi.zst</string>
-                <string>ids.zst</string>
-            </array>
-            <key>CFBundleTypeRole</key>
-            <string>Viewer</string>
-        </dict>
-        <dict>
-            <key>CFBundleTypeName</key>
-            <string>Game Boy Advance ROM</string>
-            <key>CFBundleTypeExtensions</key>
-            <array>
-                <string>gba</string>
-                <string>agb</string>
-                <string>gba.zst</string>
-                <string>agb.zst</string>
-            </array>
-            <key>CFBundleTypeRole</key>
-            <string>Viewer</string>
-            <key>LSHandlerRank</key>
-            <string>Alternate</string>
-        </dict>
-        <dict>
-            <key>CFBundleTypeName</key>
-            <string>Archive containing ROM</string>
-            <key>CFBundleTypeExtensions</key>
-            <array>
-                <key>zip</key>
-                <key>7z</key>
-                <key>rar</key>
-                <key>tar</key>
-                <key>tar.gz</key>
-                <key>tgz</key>
-                <key>tar.xz</key>
-                <key>txz</key>
-                <key>tar.bz2</key>
-                <key>tbz2</key>
-                <key>tar.lz4</key>
-                <key>tlz4</key>
-                <key>tar.zst</key>
-                <key>tzst</key>
-                <key>tar.Z</key>
-                <key>taz</key>
-                <key>tar.lz</key>
-                <key>tar.lzma</key>
-                <key>tlz</key>
-                <key>tar.lrz</key>
-                <key>tlrz</key>
-                <key>tar.lzo</key>
-                <key>tzo</key>
-            </array>
-            <key>CFBundleTypeRole</key>
-            <string>Viewer</string>
-            <key>LSHandlerRank</key>
-            <string>Alternate</string>
-        </dict>
-    </array>
-</dict>
-</plist>
--- a/res/melon.qrc
+++ b/res/melon.qrc
@ -1,7 +0,0 @@
-<!DOCTYPE RCC>
-<RCC version="1.0">
-	<qresource>
-		<file alias="melon-icon">icon/melon_256x256.png</file>
-		<file alias="melon-logo">melon384.png</file>
-	</qresource>
-</RCC>
--- a/res/melon.svg
+++ b/res/melon.svg
@ -1,80 +0,0 @@
-<svg width="1024" height="1024" xmlns="http://www.w3.org/2000/svg">
-    <path fill="#5c913b" fill-rule="evenodd" d="M357.62 154.38c56.821-56.821 85.232-85.232 118.463-94.604a132.32 132.32 0 0 1 71.834 0c33.231 9.372 61.642 37.783 118.463 94.604l203.24 203.24c56.821 56.821 85.232 85.232 94.604 118.463a132.32 132.32 0 0 1 0 71.834c-9.372 33.231-37.783 61.642-94.604 118.463L666.38 869.62c-56.821 56.821-85.232 85.232-118.463 94.604a132.319 132.319 0 0 1-71.834 0c-33.231-9.372-61.642-37.783-118.463-94.604L154.38 666.38c-56.821-56.821-85.232-85.232-94.604-118.463a132.318 132.318 0 0 1 0-71.834c9.372-33.231 37.783-61.642 94.604-118.463Z"/>
-    <filter id="a" x="0" y="0" width="1024" height="1024" filterUnits="userSpaceOnUse" primitiveUnits="userSpaceOnUse" color-interpolation-filters="sRGB">
-        <feGaussianBlur stdDeviation="18.378"/>
-        <feOffset dy="7.351" result="offsetblur"/>
-        <feFlood flood-color="#000" flood-opacity=".4"/>
-        <feComposite in2="offsetblur" operator="in"/>
-        <feMerge>
-            <feMergeNode/>
-            <feMergeNode in="SourceGraphic"/>
-        </feMerge>
-    </filter>
-    <path fill="#ffe8b6" fill-rule="evenodd" filter="url(#a)" d="M383.23 209.478c47.35-47.35 71.026-71.026 98.72-78.836a110.266 110.266 0 0 1 59.86 0c27.694 7.81 51.369 31.486 98.72 78.836L814.66 383.61c47.351 47.351 71.027 71.026 78.837 98.72a110.265 110.265 0 0 1 0 59.86c-7.81 27.694-31.486 51.37-78.837 98.72L640.53 815.042c-47.351 47.35-71.027 71.026-98.72 78.836a110.265 110.265 0 0 1-59.86 0c-27.694-7.81-51.37-31.485-98.72-78.836L209.098 640.91c-47.35-47.35-71.026-71.026-78.836-98.72a110.266 110.266 0 0 1 0-59.86c7.81-27.694 31.485-51.369 78.836-98.72Z"/>
-    <filter id="b" x="0" y="0" width="1024" height="1024" filterUnits="userSpaceOnUse" primitiveUnits="userSpaceOnUse" color-interpolation-filters="sRGB">
-        <feGaussianBlur stdDeviation="18.378"/>
-        <feOffset dy="7.351" result="offsetblur"/>
-        <feFlood flood-color="#000" flood-opacity=".4"/>
-        <feComposite in2="offsetblur" operator="in"/>
-        <feMerge>
-            <feMergeNode/>
-            <feMergeNode in="SourceGraphic"/>
-        </feMerge>
-    </filter>
-    <path fill="#dd2e44" fill-rule="evenodd" filter="url(#b)" d="M404.822 239.527c39.46-39.46 59.189-59.189 82.266-65.697a91.889 91.889 0 0 1 49.885 0c23.077 6.508 42.807 26.238 82.266 65.697l165.295 165.295c39.459 39.46 59.188 59.189 65.697 82.266a91.888 91.888 0 0 1 0 49.884c-6.509 23.078-26.238 42.808-65.697 82.266L619.239 784.534c-39.46 39.459-59.189 59.188-82.266 65.697a91.888 91.888 0 0 1-49.885 0c-23.077-6.509-42.807-26.238-82.266-65.697L239.527 619.239c-39.46-39.46-59.189-59.189-65.697-82.267a91.888 91.888 0 0 1 0-49.884c6.508-23.077 26.238-42.807 65.697-82.266Z"/>
-    <filter id="c" x="0" y="0" width="1024" height="1024" filterUnits="userSpaceOnUse" primitiveUnits="userSpaceOnUse" color-interpolation-filters="sRGB">
-        <feGaussianBlur stdDeviation="3.676"/>
-        <feOffset dx="5.198" dy="5.198" result="offsetblur"/>
-        <feFlood flood-color="#000" flood-opacity=".104"/>
-        <feComposite in2="offsetblur" operator="in"/>
-        <feMerge>
-            <feMergeNode/>
-            <feMergeNode in="SourceGraphic"/>
-        </feMerge>
-    </filter>
-    <path fill="#ff554d" fill-rule="evenodd" filter="url(#c)" d="M405.568 606.767s-14.487 11.68-31.187 15.594c-9.553 2.238-24.422 1.568-31.188-5.198l-103.96-103.96c-6.766-6.766-7.436-21.635-5.198-31.187 3.914-16.7 15.594-31.188 15.594-31.188l197.523-197.523s14.488-11.68 31.188-15.594c9.553-2.239 24.422-1.568 31.188 5.198l103.96 103.96c6.765 6.765 7.436 21.634 5.197 31.187-3.913 16.7-15.594 31.188-15.594 31.188Z"/>
-    <filter id="d" x="0" y="0" width="1024" height="1024" filterUnits="userSpaceOnUse" primitiveUnits="userSpaceOnUse" color-interpolation-filters="sRGB">
-        <feGaussianBlur stdDeviation="3.676"/>
-        <feOffset dx="5.198" dy="5.198" result="offsetblur"/>
-        <feFlood flood-color="#000" flood-opacity=".104"/>
-        <feComposite in2="offsetblur" operator="in"/>
-        <feMerge>
-            <feMergeNode/>
-            <feMergeNode in="SourceGraphic"/>
-        </feMerge>
-    </filter>
-    <path fill="#ff554d" fill-rule="evenodd" filter="url(#d)" d="M574.579 779.453s-14.488 11.68-31.188 15.594c-9.553 2.238-24.422 1.568-31.188-5.198l-103.96-103.96c-6.765-6.766-7.436-21.635-5.197-31.187 3.914-16.7 15.594-31.188 15.594-31.188L616.163 425.99s14.488-11.68 31.187-15.594c9.553-2.239 24.422-1.568 31.188 5.198l103.96 103.96c6.766 6.765 7.436 21.634 5.198 31.187-3.914 16.7-15.594 31.188-15.594 31.188Z"/>
-    <filter id="e" x="0" y="0" width="1024" height="1024" filterUnits="userSpaceOnUse" primitiveUnits="userSpaceOnUse" color-interpolation-filters="sRGB">
-        <feGaussianBlur stdDeviation="3.676"/>
-        <feOffset dx="5.198" dy="5.198" result="offsetblur"/>
-        <feFlood flood-color="#000" flood-opacity=".35"/>
-        <feComposite in2="offsetblur" operator="in"/>
-        <feMerge>
-            <feMergeNode/>
-            <feMergeNode in="SourceGraphic"/>
-        </feMerge>
-    </filter>
-    <path fill="#31352e" fill-rule="evenodd" filter="url(#e)" d="M414.022 532.978s-8.315 6.704-17.9 8.95c-5.483 1.285-14.016.9-17.9-2.983l-59.667-59.667c-3.883-3.884-4.268-12.417-2.983-17.9 2.246-9.585 8.95-17.9 8.95-17.9L437.89 330.11s8.315-6.704 17.9-8.95c5.482-1.285 14.016-.9 17.9 2.984l59.666 59.666c3.884 3.884 4.269 12.418 2.984 17.9-2.246 9.585-8.95 17.9-8.95 17.9Z"/>
-    <filter id="f" x="0" y="0" width="1024" height="1024" filterUnits="userSpaceOnUse" primitiveUnits="userSpaceOnUse" color-interpolation-filters="sRGB">
-        <feGaussianBlur stdDeviation="3.676"/>
-        <feOffset dx="5.198" dy="5.198" result="offsetblur"/>
-        <feFlood flood-color="#000" flood-opacity=".35"/>
-        <feComposite in2="offsetblur" operator="in"/>
-        <feMerge>
-            <feMergeNode/>
-            <feMergeNode in="SourceGraphic"/>
-        </feMerge>
-    </filter>
-    <path fill="#31352e" fill-rule="evenodd" filter="url(#f)" d="M677.203 544.885c7.177 7.176 18.813 7.176 25.99 0l19.954-19.955c7.177-7.176 7.177-18.813 0-25.99l-19.954-19.954c-7.177-7.176-18.813-7.176-25.99 0l-19.954 19.955c-7.177 7.176-7.177 18.813 0 25.99Z"/>
-    <filter id="g" x="0" y="0" width="1024" height="1024" filterUnits="userSpaceOnUse" primitiveUnits="userSpaceOnUse" color-interpolation-filters="sRGB">
-        <feGaussianBlur stdDeviation="3.676"/>
-        <feOffset dx="5.198" dy="5.198" result="offsetblur"/>
-        <feFlood flood-color="#000" flood-opacity=".35"/>
-        <feComposite in2="offsetblur" operator="in"/>
-        <feMerge>
-            <feMergeNode/>
-            <feMergeNode in="SourceGraphic"/>
-        </feMerge>
-    </filter>
-    <path fill="#31352e" fill-rule="evenodd" filter="url(#g)" d="M489.752 728.66c7.177 7.177 18.813 7.177 25.99 0l19.954-19.954c7.177-7.177 7.177-18.813 0-25.99l-19.954-19.954c-7.177-7.176-18.813-7.176-25.99 0l-19.954 19.955c-7.177 7.176-7.177 18.812 0 25.99Z"/>
-</svg>
--- a/res/melon384.png
+++ b/res/melon384.png
--- a/res/melon512.png
+++ b/res/melon512.png
--- a/src/ARCodeFile.cpp
+++ b/src/ARCodeFile.cpp
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -21,43 +21,32 @@
 #include "ARCodeFile.h"
 #include "Platform.h"

-namespace melonDS
-{
-using namespace Platform;

 // TODO: import codes from other sources (usrcheat.dat, ...)
 // TODO: more user-friendly error reporting


-ARCodeFile::ARCodeFile(const std::string& filename)
+ARCodeFile::ARCodeFile(const char* filename)
 {
-    Filename = filename;
+    memset(Filename, 0, sizeof(Filename));
+    strncpy(Filename, filename, 1023);
+
+    Error = false;
+
+    Categories.clear();

    if (!Load())
        Error = true;
 }

-std::vector<ARCode> ARCodeFile::GetCodes() const noexcept
+ARCodeFile::~ARCodeFile()
 {
-    if (Error)
-        return {};
-
-    std::vector<ARCode> codes;
-
-    for (const ARCodeCat& cat : Categories)
-    {
-        for (const ARCode& code : cat.Codes)
-        {
-            codes.push_back(code);
-        }
-    }
-
-    return codes;
+    Categories.clear();
 }

 bool ARCodeFile::Load()
 {
-    FileHandle* f = OpenFile(Filename, FileMode::ReadText);
+    FILE* f = Platform::OpenFile(Filename, "r");
    if (!f) return true;

    Categories.clear();
@ -69,11 +58,9 @@ bool ARCodeFile::Load()
    ARCode curcode;

    char linebuf[1024];
-    while (!IsEndOfFile(f))
+    while (!feof(f))
    {
-        if (!FileReadLine(linebuf, 1024, f))
-            break;
-
+        fgets(linebuf, 1024, f);
        linebuf[1023] = '\0';

        char* start = linebuf;
@ -91,8 +78,8 @@ bool ARCodeFile::Load()

            if (ret < 1)
            {
-                Log(LogLevel::Error, "AR: malformed CAT line: %s\n", start);
-                CloseFile(f);
+                printf("AR: malformed CAT line: %s\n", start);
+                fclose(f);
                return false;
            }

@ -102,7 +89,7 @@ bool ARCodeFile::Load()
            if (isincat) Categories.push_back(curcat);
            isincat = true;

-            curcat.Name = catname;
+            memcpy(curcat.Name, catname, 128);
            curcat.Codes.clear();
        }
        else if (!strncasecmp(start, "CODE", 4))
@ -114,24 +101,24 @@ bool ARCodeFile::Load()

            if (ret < 2)
            {
-                Log(LogLevel::Error, "AR: malformed CODE line: %s\n", start);
-                CloseFile(f);
+                printf("AR: malformed CODE line: %s\n", start);
+                fclose(f);
                return false;
            }

            if (!isincat)
            {
-                Log(LogLevel::Error, "AR: encountered CODE line with no category started\n");
-                CloseFile(f);
+                printf("AR: encountered CODE line with no category started\n");
+                fclose(f);
                return false;
            }

            if (isincode) curcat.Codes.push_back(curcode);
            isincode = true;

-            curcode.Name = codename;
+            memcpy(curcode.Name, codename, 128);
            curcode.Enabled = enable!=0;
-            curcode.Code.clear();
+            curcode.CodeLen = 0;
        }
        else
        {
@ -140,58 +127,65 @@ bool ARCodeFile::Load()

            if (ret < 2)
            {
-                Log(LogLevel::Error, "AR: malformed data line: %s\n", start);
-                CloseFile(f);
+                printf("AR: malformed data line: %s\n", start);
+                fclose(f);
                return false;
            }

            if (!isincode)
            {
-                Log(LogLevel::Error, "AR: encountered data line with no code started\n");
-                CloseFile(f);
+                printf("AR: encountered data line with no code started\n");
+                fclose(f);
                return false;
            }

-            curcode.Code.push_back(c0);
-            curcode.Code.push_back(c1);
+            if (curcode.CodeLen >= 2*64)
+            {
+                printf("AR: code too long!\n");
+                fclose(f);
+                return false;
+            }
+
+            u32 idx = curcode.CodeLen;
+            curcode.Code[idx+0] = c0;
+            curcode.Code[idx+1] = c1;
+            curcode.CodeLen += 2;
        }
    }

    if (isincode) curcat.Codes.push_back(curcode);
    if (isincat) Categories.push_back(curcat);

-    CloseFile(f);
+    fclose(f);
    return true;
 }

 bool ARCodeFile::Save()
 {
-    FileHandle* f = Platform::OpenFile(Filename, FileMode::WriteText);
+    FILE* f = Platform::OpenFile(Filename, "w");
    if (!f) return false;

    for (ARCodeCatList::iterator it = Categories.begin(); it != Categories.end(); it++)
    {
        ARCodeCat& cat = *it;

-        if (it != Categories.begin()) FileWriteFormatted(f, "\n");
-        FileWriteFormatted(f, "CAT %s\n\n", cat.Name.c_str());
+        if (it != Categories.begin()) fprintf(f, "\n");
+        fprintf(f, "CAT %s\n\n", cat.Name);

        for (ARCodeList::iterator jt = cat.Codes.begin(); jt != cat.Codes.end(); jt++)
        {
            ARCode& code = *jt;
-            FileWriteFormatted(f, "CODE %d %s\n", code.Enabled, code.Name.c_str());
+            fprintf(f, "CODE %d %s\n", code.Enabled, code.Name);

-            for (size_t i = 0; i < code.Code.size(); i+=2)
+            for (u32 i = 0; i < code.CodeLen; i+=2)
            {
-                FileWriteFormatted(f, "%08X %08X\n", code.Code[i], code.Code[i + 1]);
+                fprintf(f, "%08X %08X\n", code.Code[i], code.Code[i+1]);
            }

-            FileWriteFormatted(f, "\n");
+            fprintf(f, "\n");
        }
    }

-    CloseFile(f);
+    fclose(f);
    return true;
 }
-
-}
--- a/src/ARCodeFile.h
+++ b/src/ARCodeFile.h
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -19,27 +19,27 @@
 #ifndef ARCODEFILE_H
 #define ARCODEFILE_H

-#include <string>
 #include <list>
-#include <vector>
+
 #include "types.h"

-namespace melonDS
+typedef struct
 {
-struct ARCode
-{
-    std::string Name;
+    char Name[128];
    bool Enabled;
-    std::vector<u32> Code;
-};
+    u32 CodeLen;
+    u32 Code[2*64];
+
+} ARCode;

 typedef std::list<ARCode> ARCodeList;

-struct ARCodeCat
+typedef struct
 {
-    std::string Name;
+    char Name[128];
    ARCodeList Codes;
-};
+
+} ARCodeCat;

 typedef std::list<ARCodeCat> ARCodeCatList;

@ -47,21 +47,18 @@ typedef std::list<ARCodeCat> ARCodeCatList;
 class ARCodeFile
 {
 public:
-    ARCodeFile(const std::string& filename);
-    ~ARCodeFile() noexcept = default;
+    ARCodeFile(const char* filename);
+    ~ARCodeFile();

-    [[nodiscard]] std::vector<ARCode> GetCodes() const noexcept;
-
-    bool Error = false;
+    bool Error;

    bool Load();
    bool Save();

-    ARCodeCatList Categories {};
+    ARCodeCatList Categories;

 private:
-    std::string Filename;
+    char Filename[1024];
 };

-}
 #endif // ARCODEFILE_H
--- a/src/AREngine.cpp
+++ b/src/AREngine.cpp
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -21,34 +21,80 @@
 #include "NDS.h"
 #include "DSi.h"
 #include "AREngine.h"
-#include "Platform.h"

-namespace melonDS
+
+namespace AREngine
 {

-using Platform::Log;
-using Platform::LogLevel;
+// AR code file - frontend is responsible for managing this
+ARCodeFile* CodeFile;

-AREngine::AREngine(melonDS::NDS& nds) : NDS(nds)
+u8 (*BusRead8)(u32 addr);
+u16 (*BusRead16)(u32 addr);
+u32 (*BusRead32)(u32 addr);
+void (*BusWrite8)(u32 addr, u8 val);
+void (*BusWrite16)(u32 addr, u16 val);
+void (*BusWrite32)(u32 addr, u32 val);
+
+
+bool Init()
+{
+    CodeFile = nullptr;
+
+    return true;
+}
+
+void DeInit()
 {
 }

+void Reset()
+{
+    CodeFile = nullptr;
+
+    if (NDS::ConsoleType == 1)
+    {
+        BusRead8 = DSi::ARM7Read8;
+        BusRead16 = DSi::ARM7Read16;
+        BusRead32 = DSi::ARM7Read32;
+        BusWrite8 = DSi::ARM7Write8;
+        BusWrite16 = DSi::ARM7Write16;
+        BusWrite32 = DSi::ARM7Write32;
+    }
+    else
+    {
+        BusRead8 = NDS::ARM7Read8;
+        BusRead16 = NDS::ARM7Read16;
+        BusRead32 = NDS::ARM7Read32;
+        BusWrite8 = NDS::ARM7Write8;
+        BusWrite16 = NDS::ARM7Write16;
+        BusWrite32 = NDS::ARM7Write32;
+    }
+}
+
+
+void SetCodeFile(ARCodeFile* file)
+{
+    CodeFile = file;
+}
+
+
 #define case16(x) \
    case ((x)+0x00): case ((x)+0x01): case ((x)+0x02): case ((x)+0x03): \
    case ((x)+0x04): case ((x)+0x05): case ((x)+0x06): case ((x)+0x07): \
    case ((x)+0x08): case ((x)+0x09): case ((x)+0x0A): case ((x)+0x0B): \
    case ((x)+0x0C): case ((x)+0x0D): case ((x)+0x0E): case ((x)+0x0F)

-void AREngine::RunCheat(const ARCode& arcode)
+void RunCheat(ARCode& arcode)
 {
-    const u32* code = &arcode.Code[0];
+    u32* code = &arcode.Code[0];

    u32 offset = 0;
    u32 datareg = 0;
    u32 cond = 1;
    u32 condstack = 0;

-    const u32* loopstart = code;
+    u32* loopstart = code;
    u32 loopcount = 0;
    u32 loopcond = 1;
    u32 loopcondstack = 0;
@ -58,7 +104,7 @@ void AREngine::RunCheat(const ARCode& arcode)

    for (;;)
    {
-        if (code >= &arcode.Code[arcode.Code.size()])
+        if (code >= &arcode.Code[arcode.CodeLen])
            break;

        u32 a = *code++;
@ -83,15 +129,15 @@ void AREngine::RunCheat(const ARCode& arcode)
        switch (op)
        {
        case16(0x00): // 32-bit write
-            NDS.ARM7Write32((a & 0x0FFFFFFF) + offset, b);
+            BusWrite32((a & 0x0FFFFFFF) + offset, b);
            break;

        case16(0x10): // 16-bit write
-            NDS.ARM7Write16((a & 0x0FFFFFFF) + offset, b & 0xFFFF);
+            BusWrite16((a & 0x0FFFFFFF) + offset, b & 0xFFFF);
            break;

        case16(0x20): // 8-bit write
-            NDS.ARM7Write8((a & 0x0FFFFFFF) + offset, b & 0xFF);
+            BusWrite8((a & 0x0FFFFFFF) + offset, b & 0xFF);
            break;

        case16(0x30): // IF b > u32[a]
@ -99,9 +145,7 @@ void AREngine::RunCheat(const ARCode& arcode)
                condstack <<= 1;
                condstack |= cond;

-                u32 addr = a & 0x0FFFFFFF;
-                if (!addr) addr = offset;
-                u32 chk = NDS.ARM7Read32(addr);
+                u32 chk = BusRead32(a & 0x0FFFFFFF);

                cond = (b > chk) ? 1:0;
            }
@ -112,9 +156,7 @@ void AREngine::RunCheat(const ARCode& arcode)
                condstack <<= 1;
                condstack |= cond;

-                u32 addr = a & 0x0FFFFFFF;
-                if (!addr) addr = offset;
-                u32 chk = NDS.ARM7Read32(addr);
+                u32 chk = BusRead32(a & 0x0FFFFFFF);

                cond = (b < chk) ? 1:0;
            }
@ -125,9 +167,7 @@ void AREngine::RunCheat(const ARCode& arcode)
                condstack <<= 1;
                condstack |= cond;

-                u32 addr = a & 0x0FFFFFFF;
-                if (!addr) addr = offset;
-                u32 chk = NDS.ARM7Read32(addr);
+                u32 chk = BusRead32(a & 0x0FFFFFFF);

                cond = (b == chk) ? 1:0;
            }
@ -138,9 +178,7 @@ void AREngine::RunCheat(const ARCode& arcode)
                condstack <<= 1;
                condstack |= cond;

-                u32 addr = a & 0x0FFFFFFF;
-                if (!addr) addr = offset;
-                u32 chk = NDS.ARM7Read32(addr);
+                u32 chk = BusRead32(a & 0x0FFFFFFF);

                cond = (b != chk) ? 1:0;
            }
@ -151,9 +189,7 @@ void AREngine::RunCheat(const ARCode& arcode)
                condstack <<= 1;
                condstack |= cond;

-                u32 addr = a & 0x0FFFFFFF;
-                if (!addr) addr = offset;
-                u16 val = NDS.ARM7Read16(addr);
+                u16 val = BusRead16(a & 0x0FFFFFFF);
                u16 chk = ~(b >> 16);
                chk &= val;

@ -166,9 +202,7 @@ void AREngine::RunCheat(const ARCode& arcode)
                condstack <<= 1;
                condstack |= cond;

-                u32 addr = a & 0x0FFFFFFF;
-                if (!addr) addr = offset;
-                u16 val = NDS.ARM7Read16(addr);
+                u16 val = BusRead16(a & 0x0FFFFFFF);
                u16 chk = ~(b >> 16);
                chk &= val;

@ -181,9 +215,7 @@ void AREngine::RunCheat(const ARCode& arcode)
                condstack <<= 1;
                condstack |= cond;

-                u32 addr = a & 0x0FFFFFFF;
-                if (!addr) addr = offset;
-                u16 val = NDS.ARM7Read16(addr);
+                u16 val = BusRead16(a & 0x0FFFFFFF);
                u16 chk = ~(b >> 16);
                chk &= val;

@ -196,9 +228,7 @@ void AREngine::RunCheat(const ARCode& arcode)
                condstack <<= 1;
                condstack |= cond;

-                u32 addr = a & 0x0FFFFFFF;
-                if (!addr) addr = offset;
-                u16 val = NDS.ARM7Read16(addr);
+                u16 val = BusRead16(a & 0x0FFFFFFF);
                u16 chk = ~(b >> 16);
                chk &= val;

@ -207,7 +237,7 @@ void AREngine::RunCheat(const ARCode& arcode)
            break;

        case16(0xB0): // offset = u32[a + offset]
-            offset = NDS.ARM7Read32((a & 0x0FFFFFFF) + offset);
+            offset = BusRead32((a & 0x0FFFFFFF) + offset);
            break;

        case 0xC0: // FOR 0..b
@ -222,7 +252,7 @@ void AREngine::RunCheat(const ARCode& arcode)
            // in practice could be used for a self-modifying AR code
            // could be implemented with some hackery, but, does anything even
            // use it??
-            Log(LogLevel::Error, "AR: !! THE FUCKING C4000000 OPCODE. TELL ARISOTURA.\n");
+            printf("AR: !! THE FUCKING C4000000 OPCODE. TELL ARISOTURA.\n");
            return;

        case 0xC5: // count++ / IF (count & b.l) == b.h
@ -244,7 +274,7 @@ void AREngine::RunCheat(const ARCode& arcode)
            break;

        case 0xC6: // u32[b] = offset
-            NDS.ARM7Write32(b, offset);
+            BusWrite32(b, offset);
            break;

        case 0xD0: // ENDIF
@ -293,30 +323,30 @@ void AREngine::RunCheat(const ARCode& arcode)
            break;

        case 0xD6: // u32[b+offset] = datareg / offset += 4
-            NDS.ARM7Write32(b + offset, datareg);
+            BusWrite32(b + offset, datareg);
            offset += 4;
            break;

        case 0xD7: // u16[b+offset] = datareg / offset += 2
-            NDS.ARM7Write16(b + offset, datareg & 0xFFFF);
+            BusWrite16(b + offset, datareg & 0xFFFF);
            offset += 2;
            break;

        case 0xD8: // u8[b+offset] = datareg / offset += 1
-            NDS.ARM7Write8(b + offset, datareg & 0xFF);
+            BusWrite8(b + offset, datareg & 0xFF);
            offset += 1;
            break;

        case 0xD9: // datareg = u32[b+offset]
-            datareg = NDS.ARM7Read32(b + offset);
+            datareg = BusRead32(b + offset);
            break;

        case 0xDA: // datareg = u16[b+offset]
-            datareg = NDS.ARM7Read16(b + offset);
+            datareg = BusRead16(b + offset);
            break;

        case 0xDB: // datareg = u8[b+offset]
-            datareg = NDS.ARM7Read8(b + offset);
+            datareg = BusRead8(b + offset);
            break;

        case 0xDC: // offset += b
@ -331,8 +361,8 @@ void AREngine::RunCheat(const ARCode& arcode)
                u32 bytesleft = b;
                while (bytesleft >= 8)
                {
-                    NDS.ARM7Write32(dstaddr, *code++); dstaddr += 4;
-                    NDS.ARM7Write32(dstaddr, *code++); dstaddr += 4;
+                    BusWrite32(dstaddr, *code++); dstaddr += 4;
+                    BusWrite32(dstaddr, *code++); dstaddr += 4;
                    bytesleft -= 8;
                }
                if (bytesleft > 0)
@ -341,13 +371,13 @@ void AREngine::RunCheat(const ARCode& arcode)
                    code += 2;
                    if (bytesleft >= 4)
                    {
-                        NDS.ARM7Write32(dstaddr, *(u32*)leftover); dstaddr += 4;
+                        BusWrite32(dstaddr, *(u32*)leftover); dstaddr += 4;
                        leftover += 4;
                        bytesleft -= 4;
                    }
                    while (bytesleft > 0)
                    {
-                        NDS.ARM7Write8(dstaddr, *leftover++); dstaddr++;
+                        BusWrite8(dstaddr, *leftover++); dstaddr++;
                        bytesleft--;
                    }
                }
@ -363,14 +393,14 @@ void AREngine::RunCheat(const ARCode& arcode)
                u32 bytesleft = b;
                while (bytesleft >= 4)
                {
-                    NDS.ARM7Write32(dstaddr, NDS.ARM7Read32(srcaddr));
+                    BusWrite32(dstaddr, BusRead32(srcaddr));
                    srcaddr += 4;
                    dstaddr += 4;
                    bytesleft -= 4;
                }
                while (bytesleft > 0)
                {
-                    NDS.ARM7Write8(dstaddr, NDS.ARM7Read8(srcaddr));
+                    BusWrite8(dstaddr, BusRead8(srcaddr));
                    srcaddr++;
                    dstaddr++;
                    bytesleft--;
@ -379,20 +409,28 @@ void AREngine::RunCheat(const ARCode& arcode)
            break;

        default:
-            Log(LogLevel::Warn, "!! bad AR opcode %08X %08X\n", a, b);
+            printf("!! bad AR opcode %08X %08X\n", a, b);
            return;
        }
    }
 }

-void AREngine::RunCheats()
+void RunCheats()
 {
-    if (Cheats.empty()) return;
+    if (!CodeFile) return;

-    for (const ARCode& code : Cheats)
+    for (ARCodeCatList::iterator i = CodeFile->Categories.begin(); i != CodeFile->Categories.end(); i++)
    {
-        if (code.Enabled)
-            RunCheat(code);
+        ARCodeCat& cat = *i;
+
+        for (ARCodeList::iterator j = cat.Codes.begin(); j != cat.Codes.end(); j++)
+        {
+            ARCode& code = *j;
+
+            if (code.Enabled)
+                RunCheat(code);
+        }
    }
 }
+
 }
--- a/src/AREngine.h
+++ b/src/AREngine.h
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -19,25 +19,19 @@
 #ifndef ARENGINE_H
 #define ARENGINE_H

-#include <vector>
 #include "ARCodeFile.h"

-namespace melonDS
+namespace AREngine
 {
-class NDS;
-class AREngine
-{
-public:
-    AREngine(melonDS::NDS& nds);

-    std::vector<ARCode> Cheats {};
-private:
-    friend class ARM;
-    void RunCheats();
-    void RunCheat(const ARCode& arcode);
+bool Init();
+void DeInit();
+void Reset();

-    melonDS::NDS& NDS;
-};
+void SetCodeFile(ARCodeFile* file);
+
+void RunCheats();

 }
+
 #endif // ARENGINE_H
--- a/src/ARM.cpp
+++ b/src/ARM.cpp
--- a/src/ARM.h
+++ b/src/ARM.h
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -20,18 +20,10 @@
 #define ARM_H

 #include <algorithm>
-#include <optional>

 #include "types.h"
-#include "MemRegion.h"
-#include "MemConstants.h"
+#include "NDS.h"

-#ifdef GDBSTUB_ENABLED
-#include "debug/GdbStub.h"
-#endif
-
-namespace melonDS
-{
 inline u32 ROR(u32 x, u32 n)
 {
    return (x >> (n&0x1F)) | (x << ((32-n)&0x1F));
@ -43,33 +35,15 @@ enum
    RWFlags_ForceUser = (1<<21),
 };

-enum class CPUExecuteMode : u32
-{
-    Interpreter,
-    InterpreterGDB,
-#ifdef JIT_ENABLED
-    JIT
-#endif
-};
-
-struct GDBArgs;
-class ARMJIT;
-class GPU;
-class ARMJIT_Memory;
-class NDS;
-class Savestate;
+const u32 ITCMPhysicalSize = 0x8000;
+const u32 DTCMPhysicalSize = 0x4000;

 class ARM
-#ifdef GDBSTUB_ENABLED
-    : public Gdb::StubCallbacks
-#endif
 {
 public:
-    ARM(u32 num, bool jit, std::optional<GDBArgs> gdb, NDS& nds);
+    ARM(u32 num);
    virtual ~ARM(); // destroy shit

-    void SetGdbArgs(std::optional<GDBArgs> gdb);
-
    virtual void Reset();

    virtual void DoSavestate(Savestate* file);
@ -85,9 +59,12 @@ public:
        Halted = halt;
    }

-    void NocashPrint(u32 addr) noexcept;
+    virtual void Execute() = 0;
+#ifdef JIT_ENABLED
+    virtual void ExecuteJIT() = 0;
+#endif

-    bool CheckCondition(u32 code) const
+    bool CheckCondition(u32 code)
    {
        if (code == 0xE) return true;
        if (ConditionTable[code] & (1 << (CPSR>>28))) return true;
@ -116,19 +93,7 @@ public:
        if (v) CPSR |= 0x10000000;
    }

-    inline bool ModeIs(u32 mode) const
-    {
-        u32 cm = CPSR & 0x1f;
-        mode &= 0x1f;
-
-        if (mode == cm) return true;
-        if (mode == 0x17) return cm >= 0x14 && cm <= 0x17; // abt
-        if (mode == 0x1b) return cm >= 0x18 && cm <= 0x1b; // und
-
-        return false;
-    }
-
-    void UpdateMode(u32 oldmode, u32 newmode, bool phony = false);
+    void UpdateMode(u32 oldmode, u32 newmode);

    void TriggerIRQ();

@ -149,7 +114,6 @@ public:
    virtual void AddCycles_CDI() = 0;
    virtual void AddCycles_CD() = 0;

-    void CheckGdbIncoming();

    u32 Num;

@ -183,101 +147,75 @@ public:

    u32 ExceptionBase;

-    MemRegion CodeMem;
+    NDS::MemRegion CodeMem;

 #ifdef JIT_ENABLED
    u32 FastBlockLookupStart, FastBlockLookupSize;
    u64* FastBlockLookup;
 #endif

-    static const u32 ConditionTable[16];
-#ifdef GDBSTUB_ENABLED
-    Gdb::GdbStub GdbStub;
-#endif
-
-    melonDS::NDS& NDS;
-protected:
-    virtual u8 BusRead8(u32 addr) = 0;
-    virtual u16 BusRead16(u32 addr) = 0;
-    virtual u32 BusRead32(u32 addr) = 0;
-    virtual void BusWrite8(u32 addr, u8 val) = 0;
-    virtual void BusWrite16(u32 addr, u16 val) = 0;
-    virtual void BusWrite32(u32 addr, u32 val) = 0;
-
-#ifdef GDBSTUB_ENABLED
-    bool IsSingleStep;
-    bool BreakReq;
-    bool BreakOnStartup;
-    u16 Port;
-
-public:
-    int GetCPU() const override { return Num ? 7 : 9; }
-
-    u32 ReadReg(Gdb::Register reg) override;
-    void WriteReg(Gdb::Register reg, u32 v) override;
-    u32 ReadMem(u32 addr, int size) override;
-    void WriteMem(u32 addr, int size, u32 v) override;
-
-    void ResetGdb() override;
-    int RemoteCmd(const u8* cmd, size_t len) override;
+    static u32 ConditionTable[16];

 protected:
-#endif
-
-    void GdbCheckA();
-    void GdbCheckB();
-    void GdbCheckC();
+    u8 (*BusRead8)(u32 addr);
+    u16 (*BusRead16)(u32 addr);
+    u32 (*BusRead32)(u32 addr);
+    void (*BusWrite8)(u32 addr, u8 val);
+    void (*BusWrite16)(u32 addr, u16 val);
+    void (*BusWrite32)(u32 addr, u32 val);
 };

 class ARMv5 : public ARM
 {
 public:
-    ARMv5(melonDS::NDS& nds, std::optional<GDBArgs> gdb, bool jit);
+    ARMv5();
    ~ARMv5();

-    void Reset() override;
+    void Reset();

-    void DoSavestate(Savestate* file) override;
+    void DoSavestate(Savestate* file);

    void UpdateRegionTimings(u32 addrstart, u32 addrend);

-    void FillPipeline() override;
+    void FillPipeline();

-    void JumpTo(u32 addr, bool restorecpsr = false) override;
+    void JumpTo(u32 addr, bool restorecpsr = false);

    void PrefetchAbort();
    void DataAbort();

-    template <CPUExecuteMode mode>
    void Execute();
+#ifdef JIT_ENABLED
+    void ExecuteJIT();
+#endif

    // all code accesses are forced nonseq 32bit
    u32 CodeRead32(u32 addr, bool branch);

-    void DataRead8(u32 addr, u32* val) override;
-    void DataRead16(u32 addr, u32* val) override;
-    void DataRead32(u32 addr, u32* val) override;
-    void DataRead32S(u32 addr, u32* val) override;
-    void DataWrite8(u32 addr, u8 val) override;
-    void DataWrite16(u32 addr, u16 val) override;
-    void DataWrite32(u32 addr, u32 val) override;
-    void DataWrite32S(u32 addr, u32 val) override;
+    void DataRead8(u32 addr, u32* val);
+    void DataRead16(u32 addr, u32* val);
+    void DataRead32(u32 addr, u32* val);
+    void DataRead32S(u32 addr, u32* val);
+    void DataWrite8(u32 addr, u8 val);
+    void DataWrite16(u32 addr, u16 val);
+    void DataWrite32(u32 addr, u32 val);
+    void DataWrite32S(u32 addr, u32 val);

-    void AddCycles_C() override
+    void AddCycles_C()
    {
        // code only. always nonseq 32-bit for ARM9.
        s32 numC = (R[15] & 0x2) ? 0 : CodeCycles;
        Cycles += numC;
    }

-    void AddCycles_CI(s32 numI) override
+    void AddCycles_CI(s32 numI)
    {
        // code+internal
        s32 numC = (R[15] & 0x2) ? 0 : CodeCycles;
        Cycles += numC + numI;
    }

-    void AddCycles_CDI() override
+    void AddCycles_CDI()
    {
        // LDR/LDM cycles. ARM9 seems to skip the internal cycle there.
        // TODO: ITCM data fetches shouldn't be parallelized, they say
@ -290,7 +228,7 @@ public:
        //    Cycles += numC + numD;
    }

-    void AddCycles_CD() override
+    void AddCycles_CD()
    {
        // TODO: ITCM data fetches shouldn't be parallelized, they say
        s32 numC = (R[15] & 0x2) ? 0 : CodeCycles;
@ -302,7 +240,7 @@ public:
        //    Cycles += numC + numD;
    }

-    void GetCodeMemRegion(u32 addr, MemRegion* region);
+    void GetCodeMemRegion(u32 addr, NDS::MemRegion* region);

    void CP15Reset();
    void CP15DoSavestate(Savestate* file);
@ -320,19 +258,18 @@ public:
    void ICacheInvalidateAll();

    void CP15Write(u32 id, u32 val);
-    u32 CP15Read(u32 id) const;
+    u32 CP15Read(u32 id);

    u32 CP15Control;

    u32 RNGSeed;
-    u32 TraceProcessID;

    u32 DTCMSetting, ITCMSetting;

    // for aarch64 JIT they need to go up here
    // to be addressable by a 12-bit immediate
    u32 ITCMSize;
-    u32 DTCMBase, DTCMMask;
+    u32 DTCMBase, DTCMSize;
    s32 RegionCodeCycles;

    u8 ITCM[ITCMPhysicalSize];
@ -356,41 +293,31 @@ public:
    u8 PU_UserMap[0x100000];

    // games operate under system mode, generally
-    //#define PU_Map PU_PrivMap
-    u8* PU_Map;
+    #define PU_Map PU_PrivMap

    // code/16N/32N/32S
    u8 MemTimings[0x100000][4];

    u8* CurICacheLine;

-    bool (*GetMemRegion)(u32 addr, bool write, MemRegion* region);
-
-#ifdef GDBSTUB_ENABLED
-    u32 ReadMem(u32 addr, int size) override;
-    void WriteMem(u32 addr, int size, u32 v) override;
-#endif
-
-protected:
-    u8 BusRead8(u32 addr) override;
-    u16 BusRead16(u32 addr) override;
-    u32 BusRead32(u32 addr) override;
-    void BusWrite8(u32 addr, u8 val) override;
-    void BusWrite16(u32 addr, u16 val) override;
-    void BusWrite32(u32 addr, u32 val) override;
+    bool (*GetMemRegion)(u32 addr, bool write, NDS::MemRegion* region);
 };

 class ARMv4 : public ARM
 {
 public:
-    ARMv4(melonDS::NDS& nds, std::optional<GDBArgs> gdb, bool jit);
+    ARMv4();

-    void FillPipeline() override;
+    void Reset();

-    void JumpTo(u32 addr, bool restorecpsr = false) override;
+    void FillPipeline();
+
+    void JumpTo(u32 addr, bool restorecpsr = false);

-    template <CPUExecuteMode mode>
    void Execute();
+#ifdef JIT_ENABLED
+    void ExecuteJIT();
+#endif

    u16 CodeRead16(u32 addr)
    {
@ -402,25 +329,134 @@ public:
        return BusRead32(addr);
    }

-    void DataRead8(u32 addr, u32* val) override;
-    void DataRead16(u32 addr, u32* val) override;
-    void DataRead32(u32 addr, u32* val) override;
-    void DataRead32S(u32 addr, u32* val) override;
-    void DataWrite8(u32 addr, u8 val) override;
-    void DataWrite16(u32 addr, u16 val) override;
-    void DataWrite32(u32 addr, u32 val) override;
-    void DataWrite32S(u32 addr, u32 val) override;
-    void AddCycles_C() override;
-    void AddCycles_CI(s32 num) override;
-    void AddCycles_CDI() override;
-    void AddCycles_CD() override;
-protected:
-    u8 BusRead8(u32 addr) override;
-    u16 BusRead16(u32 addr) override;
-    u32 BusRead32(u32 addr) override;
-    void BusWrite8(u32 addr, u8 val) override;
-    void BusWrite16(u32 addr, u16 val) override;
-    void BusWrite32(u32 addr, u32 val) override;
+    void DataRead8(u32 addr, u32* val)
+    {
+        *val = BusRead8(addr);
+        DataRegion = addr;
+        DataCycles = NDS::ARM7MemTimings[addr >> 15][0];
+    }
+
+    void DataRead16(u32 addr, u32* val)
+    {
+        addr &= ~1;
+
+        *val = BusRead16(addr);
+        DataRegion = addr;
+        DataCycles = NDS::ARM7MemTimings[addr >> 15][0];
+    }
+
+    void DataRead32(u32 addr, u32* val)
+    {
+        addr &= ~3;
+
+        *val = BusRead32(addr);
+        DataRegion = addr;
+        DataCycles = NDS::ARM7MemTimings[addr >> 15][2];
+    }
+
+    void DataRead32S(u32 addr, u32* val)
+    {
+        addr &= ~3;
+
+        *val = BusRead32(addr);
+        DataCycles += NDS::ARM7MemTimings[addr >> 15][3];
+    }
+
+    void DataWrite8(u32 addr, u8 val)
+    {
+        BusWrite8(addr, val);
+        DataRegion = addr;
+        DataCycles = NDS::ARM7MemTimings[addr >> 15][0];
+    }
+
+    void DataWrite16(u32 addr, u16 val)
+    {
+        addr &= ~1;
+
+        BusWrite16(addr, val);
+        DataRegion = addr;
+        DataCycles = NDS::ARM7MemTimings[addr >> 15][0];
+    }
+
+    void DataWrite32(u32 addr, u32 val)
+    {
+        addr &= ~3;
+
+        BusWrite32(addr, val);
+        DataRegion = addr;
+        DataCycles = NDS::ARM7MemTimings[addr >> 15][2];
+    }
+
+    void DataWrite32S(u32 addr, u32 val)
+    {
+        addr &= ~3;
+
+        BusWrite32(addr, val);
+        DataCycles += NDS::ARM7MemTimings[addr >> 15][3];
+    }
+
+
+    void AddCycles_C()
+    {
+        // code only. this code fetch is sequential.
+        Cycles += NDS::ARM7MemTimings[CodeCycles][(CPSR&0x20)?1:3];
+    }
+
+    void AddCycles_CI(s32 num)
+    {
+        // code+internal. results in a nonseq code fetch.
+        Cycles += NDS::ARM7MemTimings[CodeCycles][(CPSR&0x20)?0:2] + num;
+    }
+
+    void AddCycles_CDI()
+    {
+        // LDR/LDM cycles.
+        s32 numC = NDS::ARM7MemTimings[CodeCycles][(CPSR&0x20)?0:2];
+        s32 numD = DataCycles;
+
+        if ((DataRegion >> 24) == 0x02) // mainRAM
+        {
+            if (CodeRegion == 0x02)
+                Cycles += numC + numD;
+            else
+            {
+                numC++;
+                Cycles += std::max(numC + numD - 3, std::max(numC, numD));
+            }
+        }
+        else if (CodeRegion == 0x02)
+        {
+            numD++;
+            Cycles += std::max(numC + numD - 3, std::max(numC, numD));
+        }
+        else
+        {
+            Cycles += numC + numD + 1;
+        }
+    }
+
+    void AddCycles_CD()
+    {
+        // TODO: max gain should be 5c when writing to mainRAM
+        s32 numC = NDS::ARM7MemTimings[CodeCycles][(CPSR&0x20)?0:2];
+        s32 numD = DataCycles;
+
+        if ((DataRegion >> 24) == 0x02)
+        {
+            if (CodeRegion == 0x02)
+                Cycles += numC + numD;
+            else
+                Cycles += std::max(numC + numD - 3, std::max(numC, numD));
+        }
+        else if (CodeRegion == 0x02)
+        {
+            Cycles += std::max(numC + numD - 3, std::max(numC, numD));
+        }
+        else
+        {
+            Cycles += numC + numD;
+        }
+    }
 };

 namespace ARMInterpreter
@ -430,5 +466,13 @@ void A_UNK(ARM* cpu);
 void T_UNK(ARM* cpu);

 }
+
+namespace NDS
+{
+
+extern ARMv5* ARM9;
+extern ARMv4* ARM7;
+
 }
+
 #endif // ARM_H
--- a/src/ARMInterpreter.cpp
+++ b/src/ARMInterpreter.cpp
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -22,24 +22,15 @@
 #include "ARMInterpreter_ALU.h"
 #include "ARMInterpreter_Branch.h"
 #include "ARMInterpreter_LoadStore.h"
-#include "Platform.h"

-#ifdef GDBSTUB_ENABLED
-#include "debug/GdbStub.h"
-#endif

-namespace melonDS::ARMInterpreter
+namespace ARMInterpreter
 {
-    using Platform::Log;
-    using Platform::LogLevel;


 void A_UNK(ARM* cpu)
 {
-    Log(LogLevel::Warn, "undefined ARM%d instruction %08X @ %08X\n", cpu->Num?7:9, cpu->CurInstr, cpu->R[15]-8);
-#ifdef GDBSTUB_ENABLED
-    cpu->GdbStub.Enter(cpu->GdbStub.IsConnected(), Gdb::TgtStatus::FaultInsn, cpu->R[15]-8);
-#endif
+    printf("undefined ARM%d instruction %08X @ %08X\n", cpu->Num?7:9, cpu->CurInstr, cpu->R[15]-8);
    //for (int i = 0; i < 16; i++) printf("R%d: %08X\n", i, cpu->R[i]);
    //NDS::Halt();
    u32 oldcpsr = cpu->CPSR;
@ -54,10 +45,7 @@ void A_UNK(ARM* cpu)

 void T_UNK(ARM* cpu)
 {
-    Log(LogLevel::Warn, "undefined THUMB%d instruction %04X @ %08X\n", cpu->Num?7:9, cpu->CurInstr, cpu->R[15]-4);
-#ifdef GDBSTUB_ENABLED
-    cpu->GdbStub.Enter(cpu->GdbStub.IsConnected(), Gdb::TgtStatus::FaultInsn, cpu->R[15]-4);
-#endif
+    printf("undefined THUMB%d instruction %04X @ %08X\n", cpu->Num?7:9, cpu->CurInstr, cpu->R[15]-4);
    //NDS::Halt();
    u32 oldcpsr = cpu->CPSR;
    cpu->CPSR &= ~0xBF;
@ -81,17 +69,9 @@ void A_MSR_IMM(ARM* cpu)
            case 0x11: psr = &cpu->R_FIQ[7]; break;
            case 0x12: psr = &cpu->R_IRQ[2]; break;
            case 0x13: psr = &cpu->R_SVC[2]; break;
-            case 0x14:
-            case 0x15:
-            case 0x16:
            case 0x17: psr = &cpu->R_ABT[2]; break;
-            case 0x18:
-            case 0x19:
-            case 0x1A:
            case 0x1B: psr = &cpu->R_UND[2]; break;
-            default:
-                cpu->AddCycles_C();
-                return;
+            default: printf("bad CPU mode %08X\n", cpu->CPSR); return;
        }
    }
    else
@ -112,9 +92,6 @@ void A_MSR_IMM(ARM* cpu)

    u32 val = ROR((cpu->CurInstr & 0xFF), ((cpu->CurInstr >> 7) & 0x1E));

-    // bit4 is forced to 1
-    val |= 0x00000010;
-
    *psr &= ~mask;
    *psr |= (val & mask);

@ -134,17 +111,9 @@ void A_MSR_REG(ARM* cpu)
            case 0x11: psr = &cpu->R_FIQ[7]; break;
            case 0x12: psr = &cpu->R_IRQ[2]; break;
            case 0x13: psr = &cpu->R_SVC[2]; break;
-            case 0x14:
-            case 0x15:
-            case 0x16:
            case 0x17: psr = &cpu->R_ABT[2]; break;
-            case 0x18:
-            case 0x19:
-            case 0x1A:
            case 0x1B: psr = &cpu->R_UND[2]; break;
-            default:
-                cpu->AddCycles_C();
-                return;
+            default: printf("bad CPU mode %08X\n", cpu->CPSR); return;
        }
    }
    else
@ -165,9 +134,6 @@ void A_MSR_REG(ARM* cpu)

    u32 val = cpu->R[cpu->CurInstr & 0xF];

-    // bit4 is forced to 1
-    val |= 0x00000010;
-
    *psr &= ~mask;
    *psr |= (val & mask);

@ -187,15 +153,9 @@ void A_MRS(ARM* cpu)
            case 0x11: psr = cpu->R_FIQ[7]; break;
            case 0x12: psr = cpu->R_IRQ[2]; break;
            case 0x13: psr = cpu->R_SVC[2]; break;
-            case 0x14:
-            case 0x15:
-            case 0x16:
            case 0x17: psr = cpu->R_ABT[2]; break;
-            case 0x18:
-            case 0x19:
-            case 0x1A:
            case 0x1B: psr = cpu->R_UND[2]; break;
-            default: psr = cpu->CPSR; break;
+            default: printf("bad CPU mode %08X\n", cpu->CPSR); return;
        }
    }
    else
@ -208,9 +168,6 @@ void A_MRS(ARM* cpu)

 void A_MCR(ARM* cpu)
 {
-    if ((cpu->CPSR & 0x1F) == 0x10)
-        return A_UNK(cpu);
-
    u32 cp = (cpu->CurInstr >> 8) & 0xF;
    //u32 op = (cpu->CurInstr >> 21) & 0x7;
    u32 cn = (cpu->CurInstr >> 16) & 0xF;
@ -223,11 +180,11 @@ void A_MCR(ARM* cpu)
    }
    else if (cpu->Num==1 && cp==14)
    {
-        Log(LogLevel::Debug, "MCR p14,%d,%d,%d on ARM7\n", cn, cm, cpinfo);
+        printf("MCR p14,%d,%d,%d on ARM7\n", cn, cm, cpinfo);
    }
    else
    {
-        Log(LogLevel::Warn, "bad MCR opcode p%d,%d,%d,%d on ARM%d\n", cp, cn, cm, cpinfo, cpu->Num?7:9);
+        printf("bad MCR opcode p%d,%d,%d,%d on ARM%d\n", cp, cn, cm, cpinfo, cpu->Num?7:9);
        return A_UNK(cpu); // TODO: check what kind of exception it really is
    }

@ -236,9 +193,6 @@ void A_MCR(ARM* cpu)

 void A_MRC(ARM* cpu)
 {
-    if ((cpu->CPSR & 0x1F) == 0x10)
-        return A_UNK(cpu);
-
    u32 cp = (cpu->CurInstr >> 8) & 0xF;
    //u32 op = (cpu->CurInstr >> 21) & 0x7;
    u32 cn = (cpu->CurInstr >> 16) & 0xF;
@ -251,11 +205,11 @@ void A_MRC(ARM* cpu)
    }
    else if (cpu->Num==1 && cp==14)
    {
-        Log(LogLevel::Debug, "MRC p14,%d,%d,%d on ARM7\n", cn, cm, cpinfo);
+        printf("MRC p14,%d,%d,%d on ARM7\n", cn, cm, cpinfo);
    }
    else
    {
-        Log(LogLevel::Warn, "bad MRC opcode p%d,%d,%d,%d on ARM%d\n", cp, cn, cm, cpinfo, cpu->Num?7:9);
+        printf("bad MRC opcode p%d,%d,%d,%d on ARM%d\n", cp, cn, cm, cpinfo, cpu->Num?7:9);
        return A_UNK(cpu); // TODO: check what kind of exception it really is
    }

--- a/src/ARMInterpreter.h
+++ b/src/ARMInterpreter.h
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -22,8 +22,6 @@
 #include "types.h"
 #include "ARM.h"

-namespace melonDS
-{
 namespace ARMInterpreter
 {

@ -43,5 +41,4 @@ void A_BLX_IMM(ARM* cpu); // I'm a special one look at me

 }

-}
 #endif // ARMINTERPRETER_H
--- a/src/ARMInterpreter_ALU.cpp
+++ b/src/ARMInterpreter_ALU.cpp
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -18,46 +18,18 @@

 #include <stdio.h>
 #include "ARM.h"
-#include "NDS.h"

-namespace melonDS::ARMInterpreter
+
+#define CARRY_ADD(a, b)  ((0xFFFFFFFF-a) < b)
+#define CARRY_SUB(a, b)  (a >= b)
+
+#define OVERFLOW_ADD(a, b, res)  ((!(((a) ^ (b)) & 0x80000000)) && (((a) ^ (res)) & 0x80000000))
+#define OVERFLOW_SUB(a, b, res)  ((((a) ^ (b)) & 0x80000000) && (((a) ^ (res)) & 0x80000000))
+
+
+namespace ARMInterpreter
 {

-inline bool CarryAdd(u32 a, u32 b)
-{
-    return (0xFFFFFFFF-a) < b;
-}
-
-inline bool CarrySub(u32 a, u32 b)
-{
-    return a >= b;
-}
-
-inline bool OverflowAdd(u32 a, u32 b)
-{
-    u32 res = a + b;
-    return (!((a ^ b) & 0x80000000)) && ((a ^ res) & 0x80000000);
-}
-
-inline bool OverflowSub(u32 a, u32 b)
-{
-    u32 res = a - b;
-    return ((a ^ b) & 0x80000000) && ((a ^ res) & 0x80000000);
-}
-
-inline bool OverflowAdc(u32 a, u32 b, u32 carry)
-{
-    s64 fullResult = (s64)(s32)a + (s32)b + carry;
-    u32 res = a + b + carry;
-    return (s32)res != fullResult;
-}
-
-inline bool OverflowSbc(u32 a, u32 b, u32 carry)
-{
-    s64 fullResult = (s64)(s32)a - (s32)b - carry;
-    u32 res = a - b - carry;
-    return (s32)res != fullResult;
-}

 #define LSL_IMM(x, s) \
    x <<= s;
@ -318,7 +290,7 @@ void A_##x##_REG_ROR_REG(ARM* cpu) \
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -349,7 +321,7 @@ A_IMPLEMENT_ALU_OP(AND,_S)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -380,7 +352,7 @@ A_IMPLEMENT_ALU_OP(EOR,_S)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -392,8 +364,8 @@ A_IMPLEMENT_ALU_OP(EOR,_S)
    u32 res = a - b; \
    cpu->SetNZCV(res & 0x80000000, \
                 !res, \
-                 CarrySub(a, b), \
-                 OverflowSub(a, b)); \
+                 CARRY_SUB(a, b), \
+                 OVERFLOW_SUB(a, b, res)); \
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
@ -413,7 +385,7 @@ A_IMPLEMENT_ALU_OP(SUB,)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -425,8 +397,8 @@ A_IMPLEMENT_ALU_OP(SUB,)
    u32 res = b - a; \
    cpu->SetNZCV(res & 0x80000000, \
                 !res, \
-                 CarrySub(b, a), \
-                 OverflowSub(b, a)); \
+                 CARRY_SUB(b, a), \
+                 OVERFLOW_SUB(b, a, res)); \
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
@ -446,7 +418,7 @@ A_IMPLEMENT_ALU_OP(RSB,)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -458,8 +430,8 @@ A_IMPLEMENT_ALU_OP(RSB,)
    u32 res = a + b; \
    cpu->SetNZCV(res & 0x80000000, \
                 !res, \
-                 CarryAdd(a, b), \
-                 OverflowAdd(a, b)); \
+                 CARRY_ADD(a, b), \
+                 OVERFLOW_ADD(a, b, res)); \
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
@ -479,7 +451,7 @@ A_IMPLEMENT_ALU_OP(ADD,)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -493,8 +465,8 @@ A_IMPLEMENT_ALU_OP(ADD,)
    u32 res = res_tmp + carry; \
    cpu->SetNZCV(res & 0x80000000, \
                 !res, \
-                 CarryAdd(a, b) | CarryAdd(res_tmp, carry), \
-                 OverflowAdc(a, b, carry)); \
+                 CARRY_ADD(a, b) | CARRY_ADD(res_tmp, carry), \
+                 OVERFLOW_ADD(a, b, res_tmp) | OVERFLOW_ADD(res_tmp, carry, res)); \
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
@ -514,7 +486,7 @@ A_IMPLEMENT_ALU_OP(ADC,)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -528,8 +500,8 @@ A_IMPLEMENT_ALU_OP(ADC,)
    u32 res = res_tmp - carry; \
    cpu->SetNZCV(res & 0x80000000, \
                 !res, \
-                 CarrySub(a, b) & CarrySub(res_tmp, carry), \
-                 OverflowSbc(a, b, carry)); \
+                 CARRY_SUB(a, b) & CARRY_SUB(res_tmp, carry), \
+                 OVERFLOW_SUB(a, b, res_tmp) | OVERFLOW_SUB(res_tmp, carry, res)); \
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
@ -549,7 +521,7 @@ A_IMPLEMENT_ALU_OP(SBC,)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -563,8 +535,8 @@ A_IMPLEMENT_ALU_OP(SBC,)
    u32 res = res_tmp - carry; \
    cpu->SetNZCV(res & 0x80000000, \
                 !res, \
-                 CarrySub(b, a) & CarrySub(res_tmp, carry), \
-                 OverflowSbc(b, a, carry)); \
+                 CARRY_SUB(b, a) & CARRY_SUB(res_tmp, carry), \
+                 OVERFLOW_SUB(b, a, res_tmp) | OVERFLOW_SUB(res_tmp, carry, res)); \
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
@ -603,8 +575,8 @@ A_IMPLEMENT_ALU_TEST(TEQ,_S)
    u32 res = a - b; \
    cpu->SetNZCV(res & 0x80000000, \
                 !res, \
-                 CarrySub(a, b), \
-                 OverflowSub(a, b)); \
+                 CARRY_SUB(a, b), \
+                 OVERFLOW_SUB(a, b, res)); \
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C();

 A_IMPLEMENT_ALU_TEST(CMP,)
@ -615,8 +587,8 @@ A_IMPLEMENT_ALU_TEST(CMP,)
    u32 res = a + b; \
    cpu->SetNZCV(res & 0x80000000, \
                 !res, \
-                 CarryAdd(a, b), \
-                 OverflowAdd(a, b)); \
+                 CARRY_ADD(a, b), \
+                 OVERFLOW_ADD(a, b, res)); \
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C();

 A_IMPLEMENT_ALU_TEST(CMN,)
@ -628,7 +600,7 @@ A_IMPLEMENT_ALU_TEST(CMN,)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -657,7 +629,7 @@ A_IMPLEMENT_ALU_OP(ORR,_S)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(b & ~1); \
+        cpu->JumpTo(b); \
    } \
    else \
    { \
@ -689,11 +661,8 @@ void A_MOV_REG_LSL_IMM_DBG(ARM* cpu)
        (cpu->NextInstr[0] & 0xFF000000) == 0xEA000000 && // branch
        (cpu->NextInstr[1] & 0xFFFF) == 0x6464)
    {
-        // GBATek says the two bytes after the 2nd ID are _reserved_ for flags
-        // but since they serve no purpose ATTOW, we can skip them
-        u32 addr = cpu->R[15] + 4; // Skip 2nd ID and flags
-        // TODO: Pass flags to NocashPrint
-        cpu->NDS.NocashPrint(cpu->Num, addr);
+        u32 addr = cpu->R[15] + 2;
+        NDS::NocashPrint(cpu->Num, addr);
    }
 }

@ -704,7 +673,7 @@ void A_MOV_REG_LSL_IMM_DBG(ARM* cpu)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(res & ~1); \
+        cpu->JumpTo(res); \
    } \
    else \
    { \
@ -734,7 +703,7 @@ A_IMPLEMENT_ALU_OP(BIC,_S)
    if (c) cpu->AddCycles_CI(c); else cpu->AddCycles_C(); \
    if (((cpu->CurInstr>>12) & 0xF) == 15) \
    { \
-        cpu->JumpTo(b & ~1); \
+        cpu->JumpTo(b); \
    } \
    else \
    { \
@ -961,7 +930,7 @@ void A_SMLAxy(ARM* cpu)
    u32 res = res_mul + rn;

    cpu->R[(cpu->CurInstr >> 16) & 0xF] = res;
-    if (OverflowAdd(res_mul, rn))
+    if (OVERFLOW_ADD(res_mul, rn, res))
        cpu->CPSR |= 0x08000000;

    cpu->AddCycles_C(); // TODO: interlock??
@ -982,7 +951,7 @@ void A_SMLAWy(ARM* cpu)
    u32 res = res_mul + rn;

    cpu->R[(cpu->CurInstr >> 16) & 0xF] = res;
-    if (OverflowAdd(res_mul, rn))
+    if (OVERFLOW_ADD(res_mul, rn, res))
        cpu->CPSR |= 0x08000000;

    cpu->AddCycles_C(); // TODO: interlock??
@ -1079,7 +1048,7 @@ void A_QADD(ARM* cpu)
    u32 rn = cpu->R[(cpu->CurInstr >> 16) & 0xF];

    u32 res = rm + rn;
-    if (OverflowAdd(rm, rn))
+    if (OVERFLOW_ADD(rm, rn, res))
    {
        res = (res & 0x80000000) ? 0x7FFFFFFF : 0x80000000;
        cpu->CPSR |= 0x08000000;
@ -1097,7 +1066,7 @@ void A_QSUB(ARM* cpu)
    u32 rn = cpu->R[(cpu->CurInstr >> 16) & 0xF];

    u32 res = rm - rn;
-    if (OverflowSub(rm, rn))
+    if (OVERFLOW_SUB(rm, rn, res))
    {
        res = (res & 0x80000000) ? 0x7FFFFFFF : 0x80000000;
        cpu->CPSR |= 0x08000000;
@ -1114,7 +1083,7 @@ void A_QDADD(ARM* cpu)
    u32 rm = cpu->R[cpu->CurInstr & 0xF];
    u32 rn = cpu->R[(cpu->CurInstr >> 16) & 0xF];

-    if (OverflowAdd(rn, rn))
+    if (OVERFLOW_ADD(rn, rn, rn<<1))
    {
        rn = (rn & 0x80000000) ? 0x80000000 : 0x7FFFFFFF;
        cpu->CPSR |= 0x08000000; // CHECKME
@ -1123,7 +1092,7 @@ void A_QDADD(ARM* cpu)
        rn <<= 1;

    u32 res = rm + rn;
-    if (OverflowAdd(rm, rn))
+    if (OVERFLOW_ADD(rm, rn, res))
    {
        res = (res & 0x80000000) ? 0x7FFFFFFF : 0x80000000;
        cpu->CPSR |= 0x08000000;
@ -1140,7 +1109,7 @@ void A_QDSUB(ARM* cpu)
    u32 rm = cpu->R[cpu->CurInstr & 0xF];
    u32 rn = cpu->R[(cpu->CurInstr >> 16) & 0xF];

-    if (OverflowAdd(rn, rn))
+    if (OVERFLOW_ADD(rn, rn, rn<<1))
    {
        rn = (rn & 0x80000000) ? 0x80000000 : 0x7FFFFFFF;
        cpu->CPSR |= 0x08000000; // CHECKME
@ -1149,7 +1118,7 @@ void A_QDSUB(ARM* cpu)
        rn <<= 1;

    u32 res = rm - rn;
-    if (OverflowSub(rm, rn))
+    if (OVERFLOW_SUB(rm, rn, res))
    {
        res = (res & 0x80000000) ? 0x7FFFFFFF : 0x80000000;
        cpu->CPSR |= 0x08000000;
@ -1206,8 +1175,8 @@ void T_ADD_REG_(ARM* cpu)
    cpu->R[cpu->CurInstr & 0x7] = res;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarryAdd(a, b),
-                 OverflowAdd(a, b));
+                 CARRY_ADD(a, b),
+                 OVERFLOW_ADD(a, b, res));
    cpu->AddCycles_C();
 }

@ -1219,8 +1188,8 @@ void T_SUB_REG_(ARM* cpu)
    cpu->R[cpu->CurInstr & 0x7] = res;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarrySub(a, b),
-                 OverflowSub(a, b));
+                 CARRY_SUB(a, b),
+                 OVERFLOW_SUB(a, b, res));
    cpu->AddCycles_C();
 }

@ -1232,8 +1201,8 @@ void T_ADD_IMM_(ARM* cpu)
    cpu->R[cpu->CurInstr & 0x7] = res;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarryAdd(a, b),
-                 OverflowAdd(a, b));
+                 CARRY_ADD(a, b),
+                 OVERFLOW_ADD(a, b, res));
    cpu->AddCycles_C();
 }

@ -1245,8 +1214,8 @@ void T_SUB_IMM_(ARM* cpu)
    cpu->R[cpu->CurInstr & 0x7] = res;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarrySub(a, b),
-                 OverflowSub(a, b));
+                 CARRY_SUB(a, b),
+                 OVERFLOW_SUB(a, b, res));
    cpu->AddCycles_C();
 }

@ -1266,8 +1235,8 @@ void T_CMP_IMM(ARM* cpu)
    u32 res = a - b;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarrySub(a, b),
-                 OverflowSub(a, b));
+                 CARRY_SUB(a, b),
+                 OVERFLOW_SUB(a, b, res));
    cpu->AddCycles_C();
 }

@ -1279,8 +1248,8 @@ void T_ADD_IMM(ARM* cpu)
    cpu->R[(cpu->CurInstr >> 8) & 0x7] = res;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarryAdd(a, b),
-                 OverflowAdd(a, b));
+                 CARRY_ADD(a, b),
+                 OVERFLOW_ADD(a, b, res));
    cpu->AddCycles_C();
 }

@ -1292,8 +1261,8 @@ void T_SUB_IMM(ARM* cpu)
    cpu->R[(cpu->CurInstr >> 8) & 0x7] = res;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarrySub(a, b),
-                 OverflowSub(a, b));
+                 CARRY_SUB(a, b),
+                 OVERFLOW_SUB(a, b, res));
    cpu->AddCycles_C();
 }

@ -1363,8 +1332,8 @@ void T_ADC_REG(ARM* cpu)
    cpu->R[cpu->CurInstr & 0x7] = res;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarryAdd(a, b) | CarryAdd(res_tmp, carry),
-                 OverflowAdc(a, b, carry));
+                 CARRY_ADD(a, b) | CARRY_ADD(res_tmp, carry),
+                 OVERFLOW_ADD(a, b, res_tmp) | OVERFLOW_ADD(res_tmp, carry, res));
    cpu->AddCycles_C();
 }

@ -1378,8 +1347,8 @@ void T_SBC_REG(ARM* cpu)
    cpu->R[cpu->CurInstr & 0x7] = res;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarrySub(a, b) & CarrySub(res_tmp, carry),
-                 OverflowSbc(a, b, carry));
+                 CARRY_SUB(a, b) & CARRY_SUB(res_tmp, carry),
+                 OVERFLOW_SUB(a, b, res_tmp) | OVERFLOW_SUB(res_tmp, carry, res));
    cpu->AddCycles_C();
 }

@ -1411,8 +1380,8 @@ void T_NEG_REG(ARM* cpu)
    cpu->R[cpu->CurInstr & 0x7] = res;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarrySub(0, b),
-                 OverflowSub(0, b));
+                 CARRY_SUB(0, b),
+                 OVERFLOW_SUB(0, b, res));
    cpu->AddCycles_C();
 }

@ -1423,8 +1392,8 @@ void T_CMP_REG(ARM* cpu)
    u32 res = a - b;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarrySub(a, b),
-                 OverflowSub(a, b));
+                 CARRY_SUB(a, b),
+                 OVERFLOW_SUB(a, b, res));
    cpu->AddCycles_C();
 }

@ -1435,8 +1404,8 @@ void T_CMN_REG(ARM* cpu)
    u32 res = a + b;
    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarryAdd(a, b),
-                 OverflowAdd(a, b));
+                 CARRY_ADD(a, b),
+                 OVERFLOW_ADD(a, b, res));
    cpu->AddCycles_C();
 }

@ -1532,8 +1501,8 @@ void T_CMP_HIREG(ARM* cpu)

    cpu->SetNZCV(res & 0x80000000,
                 !res,
-                 CarrySub(a, b),
-                 OverflowSub(a, b));
+                 CARRY_SUB(a, b),
+                 OVERFLOW_SUB(a, b, res));
    cpu->AddCycles_C();
 }

@ -1558,11 +1527,8 @@ void T_MOV_HIREG(ARM* cpu)
        (cpu->NextInstr[0] & 0xF800) == 0xE000 && // branch
        (cpu->NextInstr[1] & 0xFFFF) == 0x6464)
    {
-        // GBATek says the two bytes after the 2nd ID are _reserved_ for flags
-        // but since they serve no purpose ATTOW, we can skip them
-        u32 addr = cpu->R[15] + 4; // Skip 2nd ID and flags
-        // TODO: Pass flags to NocashPrint
-        cpu->NDS.NocashPrint(cpu->Num, addr);
+        u32 addr = cpu->R[15] + 2;
+        NDS::NocashPrint(cpu->Num, addr);
    }
 }

--- a/src/ARMInterpreter_ALU.h
+++ b/src/ARMInterpreter_ALU.h
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -19,8 +19,6 @@
 #ifndef ARMINTERPRETER_ALU_H
 #define ARMINTERPRETER_ALU_H

-namespace melonDS
-{
 namespace ARMInterpreter
 {

@ -136,5 +134,4 @@ void T_ADD_SP(ARM* cpu);

 }

-}
 #endif
--- a/src/ARMInterpreter_Branch.cpp
+++ b/src/ARMInterpreter_Branch.cpp
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -16,13 +16,12 @@
    with melonDS. If not, see http://www.gnu.org/licenses/.
 */

+#include <stdio.h>
 #include "ARM.h"
-#include "Platform.h"

-namespace melonDS::ARMInterpreter
+
+namespace ARMInterpreter
 {
-using Platform::Log;
-using Platform::LogLevel;


 void A_B(ARM* cpu)
@ -80,7 +79,7 @@ void T_BLX_REG(ARM* cpu)
 {
    if (cpu->Num==1)
    {
-        Log(LogLevel::Warn, "!! THUMB BLX_REG ON ARM7\n");
+        printf("!! THUMB BLX_REG ON ARM7\n");
        return;
    }

--- a/src/ARMInterpreter_Branch.h
+++ b/src/ARMInterpreter_Branch.h
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -19,8 +19,6 @@
 #ifndef ARMINTERPRETER_BRANCH_H
 #define ARMINTERPRETER_BRANCH_H

-namespace melonDS
-{
 namespace ARMInterpreter
 {

@ -38,5 +36,4 @@ void T_BL_LONG_2(ARM* cpu);

 }

-}
 #endif
--- a/src/ARMInterpreter_LoadStore.cpp
+++ b/src/ARMInterpreter_LoadStore.cpp
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -20,7 +20,7 @@
 #include "ARM.h"


-namespace melonDS::ARMInterpreter
+namespace ARMInterpreter
 {


@ -62,20 +62,14 @@ namespace melonDS::ARMInterpreter

 #define A_STR \
    offset += cpu->R[(cpu->CurInstr>>16) & 0xF]; \
-    u32 storeval = cpu->R[(cpu->CurInstr>>12) & 0xF]; \
-    if (((cpu->CurInstr>>12) & 0xF) == 0xF) \
-        storeval += 4; \
-    cpu->DataWrite32(offset, storeval); \
+    cpu->DataWrite32(offset, cpu->R[(cpu->CurInstr>>12) & 0xF]); \
    if (cpu->CurInstr & (1<<21)) cpu->R[(cpu->CurInstr>>16) & 0xF] = offset; \
    cpu->AddCycles_CD();

 // TODO: user mode (bit21)
 #define A_STR_POST \
    u32 addr = cpu->R[(cpu->CurInstr>>16) & 0xF]; \
-    u32 storeval = cpu->R[(cpu->CurInstr>>12) & 0xF]; \
-    if (((cpu->CurInstr>>12) & 0xF) == 0xF) \
-        storeval += 4; \
-    cpu->DataWrite32(addr, storeval); \
+    cpu->DataWrite32(addr, cpu->R[(cpu->CurInstr>>12) & 0xF]); \
    cpu->R[(cpu->CurInstr>>16) & 0xF] += offset; \
    cpu->AddCycles_CD();

@ -416,7 +410,7 @@ void A_LDM(ARM* cpu)
    }

    if ((cpu->CurInstr & (1<<22)) && !(cpu->CurInstr & (1<<15)))
-        cpu->UpdateMode(cpu->CPSR, (cpu->CPSR&~0x1F)|0x10, true);
+        cpu->UpdateMode(cpu->CPSR, (cpu->CPSR&~0x1F)|0x10);

    for (int i = 0; i < 15; i++)
    {
@ -430,9 +424,9 @@ void A_LDM(ARM* cpu)
        }
    }

-    u32 pc = 0;
    if (cpu->CurInstr & (1<<15))
    {
+        u32 pc;
        if (preinc) base += 4;
        if (first) cpu->DataRead32 (base, &pc);
        else       cpu->DataRead32S(base, &pc);
@ -440,8 +434,13 @@ void A_LDM(ARM* cpu)

        if (cpu->Num == 1)
            pc &= ~0x1;
+
+        cpu->JumpTo(pc, cpu->CurInstr & (1<<22));
    }

+    if ((cpu->CurInstr & (1<<22)) && !(cpu->CurInstr & (1<<15)))
+        cpu->UpdateMode((cpu->CPSR&~0x1F)|0x10, cpu->CPSR);
+
    if (cpu->CurInstr & (1<<21))
    {
        // post writeback
@ -461,12 +460,6 @@ void A_LDM(ARM* cpu)
            cpu->R[baseid] = wbbase;
    }

-    if ((cpu->CurInstr & (1<<22)) && !(cpu->CurInstr & (1<<15)))
-        cpu->UpdateMode((cpu->CPSR&~0x1F)|0x10, cpu->CPSR, true);
-
-    if (cpu->CurInstr & (1<<15))
-        cpu->JumpTo(pc, cpu->CurInstr & (1<<22));
-
    cpu->AddCycles_CDI();
 }

@ -501,7 +494,7 @@ void A_STM(ARM* cpu)
        else if (mode != 0x10 && mode != 0x1F)
            isbanked = (baseid >= 13 && baseid < 15);

-        cpu->UpdateMode(cpu->CPSR, (cpu->CPSR&~0x1F)|0x10, true);
+        cpu->UpdateMode(cpu->CPSR, (cpu->CPSR&~0x1F)|0x10);
    }

    for (u32 i = 0; i < 16; i++)
@ -527,7 +520,7 @@ void A_STM(ARM* cpu)
    }

    if (cpu->CurInstr & (1<<22))
-        cpu->UpdateMode((cpu->CPSR&~0x1F)|0x10, cpu->CPSR, true);
+        cpu->UpdateMode((cpu->CPSR&~0x1F)|0x10, cpu->CPSR);

    if ((cpu->CurInstr & (1<<23)) && (cpu->CurInstr & (1<<21)))
        cpu->R[baseid] = base;
--- a/src/ARMInterpreter_LoadStore.h
+++ b/src/ARMInterpreter_LoadStore.h
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

@ -19,7 +19,7 @@
 #ifndef ARMINTERPRETER_LOADSTORE_H
 #define ARMINTERPRETER_LOADSTORE_H

-namespace melonDS::ARMInterpreter
+namespace ARMInterpreter
 {

 #define A_PROTO_WB_LDRSTR(x) \
--- a/src/ARMJIT.cpp
+++ b/src/ARMJIT.cpp
@ -1,23 +1,5 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARMJIT.h"
-#include "ARMJIT_Memory.h"
+
 #include <string.h>
 #include <assert.h>
 #include <unordered_map>
@ -25,12 +7,11 @@
 #define XXH_STATIC_LINKING_ONLY
 #include "xxhash/xxhash.h"

-#include "Platform.h"
+#include "Config.h"

 #include "ARMJIT_Internal.h"
 #include "ARMJIT_Memory.h"
 #include "ARMJIT_Compiler.h"
-#include "ARMJIT_Global.h"

 #include "ARMInterpreter_ALU.h"
 #include "ARMInterpreter_LoadStore.h"
@ -43,52 +24,130 @@
 #include "SPU.h"
 #include "Wifi.h"
 #include "NDSCart.h"
-#include "Platform.h"
+
 #include "ARMJIT_x64/ARMJIT_Offsets.h"
+static_assert(offsetof(ARM, CPSR) == ARM_CPSR_offset);
+static_assert(offsetof(ARM, Cycles) == ARM_Cycles_offset);
+static_assert(offsetof(ARM, StopExecution) == ARM_StopExecution_offset);

-namespace melonDS
+namespace ARMJIT
 {
-using Platform::Log;
-using Platform::LogLevel;
-
-static_assert(offsetof(ARM, CPSR) == ARM_CPSR_offset, "");
-static_assert(offsetof(ARM, Cycles) == ARM_Cycles_offset, "");
-static_assert(offsetof(ARM, StopExecution) == ARM_StopExecution_offset, "");
-

 #define JIT_DEBUGPRINT(msg, ...)
-//#define JIT_DEBUGPRINT(msg, ...) Platform::Log(Platform::LogLevel::Debug, msg, ## __VA_ARGS__)
+//#define JIT_DEBUGPRINT(msg, ...) printf(msg, ## __VA_ARGS__)
+
+Compiler* JITCompiler;
+
+
+std::unordered_map<u32, JitBlock*> JitBlocks9;
+std::unordered_map<u32, JitBlock*> JitBlocks7;
+
+std::unordered_map<u32, JitBlock*> RestoreCandidates;
+
+TinyVector<u32> InvalidLiterals;
+
+AddressRange CodeIndexITCM[ITCMPhysicalSize / 512];
+AddressRange CodeIndexMainRAM[NDS::MainRAMMaxSize / 512];
+AddressRange CodeIndexSWRAM[NDS::SharedWRAMSize / 512];
+AddressRange CodeIndexVRAM[0x100000 / 512];
+AddressRange CodeIndexARM9BIOS[sizeof(NDS::ARM9BIOS) / 512];
+AddressRange CodeIndexARM7BIOS[sizeof(NDS::ARM7BIOS) / 512];
+AddressRange CodeIndexARM7WRAM[NDS::ARM7WRAMSize / 512];
+AddressRange CodeIndexARM7WVRAM[0x40000 / 512];
+AddressRange CodeIndexBIOS9DSi[0x10000 / 512];
+AddressRange CodeIndexBIOS7DSi[0x10000 / 512];
+AddressRange CodeIndexNWRAM_A[DSi::NWRAMSize / 512];
+AddressRange CodeIndexNWRAM_B[DSi::NWRAMSize / 512];
+AddressRange CodeIndexNWRAM_C[DSi::NWRAMSize / 512];
+
+u64 FastBlockLookupITCM[ITCMPhysicalSize / 2];
+u64 FastBlockLookupMainRAM[NDS::MainRAMMaxSize / 2];
+u64 FastBlockLookupSWRAM[NDS::SharedWRAMSize / 2];
+u64 FastBlockLookupVRAM[0x100000 / 2];
+u64 FastBlockLookupARM9BIOS[sizeof(NDS::ARM9BIOS) / 2];
+u64 FastBlockLookupARM7BIOS[sizeof(NDS::ARM7BIOS) / 2];
+u64 FastBlockLookupARM7WRAM[NDS::ARM7WRAMSize / 2];
+u64 FastBlockLookupARM7WVRAM[0x40000 / 2];
+u64 FastBlockLookupBIOS9DSi[0x10000 / 2];
+u64 FastBlockLookupBIOS7DSi[0x10000 / 2];
+u64 FastBlockLookupNWRAM_A[DSi::NWRAMSize / 2];
+u64 FastBlockLookupNWRAM_B[DSi::NWRAMSize / 2];
+u64 FastBlockLookupNWRAM_C[DSi::NWRAMSize / 2];

 const u32 CodeRegionSizes[ARMJIT_Memory::memregions_Count] =
 {
    0,
    ITCMPhysicalSize,
    0,
-    ARM9BIOSSize,
-    MainRAMMaxSize,
-    SharedWRAMSize,
+    sizeof(NDS::ARM9BIOS),
+    NDS::MainRAMMaxSize,
+    NDS::SharedWRAMSize,
    0,
    0x100000,
-    ARM7BIOSSize,
-    ARM7WRAMSize,
+    sizeof(NDS::ARM7BIOS),
+    NDS::ARM7WRAMSize,
    0,
    0,
    0x40000,
    0x10000,
    0x10000,
-    NWRAMSize,
-    NWRAMSize,
-    NWRAMSize,
+    DSi::NWRAMSize,
+    DSi::NWRAMSize,
+    DSi::NWRAMSize,
 };

-u32 ARMJIT::LocaliseCodeAddress(u32 num, u32 addr) const noexcept
+AddressRange* const CodeMemRegions[ARMJIT_Memory::memregions_Count] =
+{
+    NULL,
+    CodeIndexITCM,
+    NULL,
+    CodeIndexARM9BIOS,
+    CodeIndexMainRAM,
+    CodeIndexSWRAM,
+    NULL,
+    CodeIndexVRAM,
+    CodeIndexARM7BIOS,
+    CodeIndexARM7WRAM,
+    NULL,
+    NULL,
+    CodeIndexARM7WVRAM,
+    CodeIndexBIOS9DSi,
+    CodeIndexBIOS7DSi,
+    CodeIndexNWRAM_A,
+    CodeIndexNWRAM_B,
+    CodeIndexNWRAM_C
+};
+
+u64* const FastBlockLookupRegions[ARMJIT_Memory::memregions_Count] =
+{
+    NULL,
+    FastBlockLookupITCM,
+    NULL,
+    FastBlockLookupARM9BIOS,
+    FastBlockLookupMainRAM,
+    FastBlockLookupSWRAM,
+    NULL,
+    FastBlockLookupVRAM,
+    FastBlockLookupARM7BIOS,
+    FastBlockLookupARM7WRAM,
+    NULL,
+    NULL,
+    FastBlockLookupARM7WVRAM,
+    FastBlockLookupBIOS9DSi,
+    FastBlockLookupBIOS7DSi,
+    FastBlockLookupNWRAM_A,
+    FastBlockLookupNWRAM_B,
+    FastBlockLookupNWRAM_C
+};
+
+u32 LocaliseCodeAddress(u32 num, u32 addr)
 {
    int region = num == 0
-        ? Memory.ClassifyAddress9(addr)
-        : Memory.ClassifyAddress7(addr);
+        ? ARMJIT_Memory::ClassifyAddress9(addr)
+        : ARMJIT_Memory::ClassifyAddress7(addr);

    if (CodeMemRegions[region])
-        return Memory.LocaliseAddress(region, num, addr);
+        return ARMJIT_Memory::LocaliseAddress(region, num, addr);
    return 0;
 }

@ -101,34 +160,14 @@ T SlowRead9(u32 addr, ARMv5* cpu)
    T val;
    if (addr < cpu->ITCMSize)
        val = *(T*)&cpu->ITCM[addr & 0x7FFF];
-    else if ((addr & cpu->DTCMMask) == cpu->DTCMBase)
-        val = *(T*)&cpu->DTCM[addr & 0x3FFF];
+    else if (addr >= cpu->DTCMBase && addr < (cpu->DTCMBase + cpu->DTCMSize))
+        val = *(T*)&cpu->DTCM[(addr - cpu->DTCMBase) & 0x3FFF];
    else if (std::is_same<T, u32>::value)
-        val = NDS::Current->ARM9Read32(addr);
+        val = (ConsoleType == 0 ? NDS::ARM9Read32 : DSi::ARM9Read32)(addr);
    else if (std::is_same<T, u16>::value)
-        val = NDS::Current->ARM9Read16(addr);
+        val = (ConsoleType == 0 ? NDS::ARM9Read16 : DSi::ARM9Read16)(addr);
    else
-        val = NDS::Current->ARM9Read8(addr);
-
-    if (std::is_same<T, u32>::value)
-        return ROR(val, offset << 3);
-    else
-        return val;
-}
-
-template <typename T, int ConsoleType>
-T SlowRead7(u32 addr)
-{
-    u32 offset = addr & 0x3;
-    addr &= ~(sizeof(T) - 1);
-
-    T val;
-    if (std::is_same<T, u32>::value)
-        val = NDS::Current->ARM7Read32(addr);
-    else if (std::is_same<T, u16>::value)
-        val = NDS::Current->ARM7Read16(addr);
-    else
-        val = NDS::Current->ARM7Read8(addr);
+        val = (ConsoleType == 0 ? NDS::ARM9Read8 : DSi::ARM9Read8)(addr);

    if (std::is_same<T, u32>::value)
        return ROR(val, offset << 3);
@ -143,45 +182,65 @@ void SlowWrite9(u32 addr, ARMv5* cpu, u32 val)

    if (addr < cpu->ITCMSize)
    {
-        cpu->NDS.JIT.CheckAndInvalidate<0, ARMJIT_Memory::memregion_ITCM>(addr);
+        CheckAndInvalidate<0, ARMJIT_Memory::memregion_ITCM>(addr);
        *(T*)&cpu->ITCM[addr & 0x7FFF] = val;
    }
-    else if ((addr & cpu->DTCMMask) == cpu->DTCMBase)
+    else if (addr >= cpu->DTCMBase && addr < (cpu->DTCMBase + cpu->DTCMSize))
    {
-        *(T*)&cpu->DTCM[addr & 0x3FFF] = val;
+        *(T*)&cpu->DTCM[(addr - cpu->DTCMBase) & 0x3FFF] = val;
    }
    else if (std::is_same<T, u32>::value)
    {
-        NDS::Current->ARM9Write32(addr, val);
+        (ConsoleType == 0 ? NDS::ARM9Write32 : DSi::ARM9Write32)(addr, val);
    }
    else if (std::is_same<T, u16>::value)
    {
-        NDS::Current->ARM9Write16(addr, val);
+        (ConsoleType == 0 ? NDS::ARM9Write16 : DSi::ARM9Write16)(addr, val);
    }
    else
    {
-        NDS::Current->ARM9Write8(addr, val);
+        (ConsoleType == 0 ? NDS::ARM9Write8 : DSi::ARM9Write8)(addr, val);
    }
 }

+template <typename T, int ConsoleType>
+T SlowRead7(u32 addr)
+{
+    u32 offset = addr & 0x3;
+    addr &= ~(sizeof(T) - 1);
+
+    T val;
+    if (std::is_same<T, u32>::value)
+        val = (ConsoleType == 0 ? NDS::ARM7Read32 : DSi::ARM7Read32)(addr);
+    else if (std::is_same<T, u16>::value)
+        val = (ConsoleType == 0 ? NDS::ARM7Read16 : DSi::ARM7Read16)(addr);
+    else
+        val = (ConsoleType == 0 ? NDS::ARM7Read8 : DSi::ARM7Read8)(addr);
+
+    if (std::is_same<T, u32>::value)
+        return ROR(val, offset << 3);
+    else
+        return val;
+}
+
 template <typename T, int ConsoleType>
 void SlowWrite7(u32 addr, u32 val)
 {
    addr &= ~(sizeof(T) - 1);

    if (std::is_same<T, u32>::value)
-        NDS::Current->ARM7Write32(addr, val);
+        (ConsoleType == 0 ? NDS::ARM7Write32 : DSi::ARM7Write32)(addr, val);
    else if (std::is_same<T, u16>::value)
-        NDS::Current->ARM7Write16(addr, val);
+        (ConsoleType == 0 ? NDS::ARM7Write16 : DSi::ARM7Write16)(addr, val);
    else
-        NDS::Current->ARM7Write8(addr, val);
+        (ConsoleType == 0 ? NDS::ARM7Write8 : DSi::ARM7Write8)(addr, val);
 }

 template <bool Write, int ConsoleType>
 void SlowBlockTransfer9(u32 addr, u64* data, u32 num, ARMv5* cpu)
 {
    addr &= ~0x3;
-    for (u32 i = 0; i < num; i++)
+    for (int i = 0; i < num; i++)
    {
        if (Write)
            SlowWrite9<u32, ConsoleType>(addr, cpu, data[i]);
@ -195,7 +254,7 @@ template <bool Write, int ConsoleType>
 void SlowBlockTransfer7(u32 addr, u64* data, u32 num)
 {
    addr &= ~0x3;
-    for (u32 i = 0; i < num; i++)
+    for (int i = 0; i < num; i++)
    {
        if (Write)
            SlowWrite7<u32, ConsoleType>(addr, data[i]);
@ -230,18 +289,26 @@ void SlowBlockTransfer7(u32 addr, u64* data, u32 num)
 INSTANTIATE_SLOWMEM(0)
 INSTANTIATE_SLOWMEM(1)

-ARMJIT::~ARMJIT() noexcept
+void Init()
 {
-    JitEnableWrite();
-    ResetBlockCache();
+    JITCompiler = new Compiler();
+
+    ARMJIT_Memory::Init();
 }

-void ARMJIT::Reset() noexcept
+void DeInit()
+{
+    ResetBlockCache();
+    ARMJIT_Memory::DeInit();
+
+    delete JITCompiler;
+}
+
+void Reset()
 {
-    JitEnableWrite();
    ResetBlockCache();

-    Memory.Reset();
+    ARMJIT_Memory::Reset();
 }

 void FloodFillSetFlags(FetchedInstr instrs[], int start, u8 flags)
@ -289,7 +356,7 @@ bool DecodeLiteral(bool thumb, const FetchedInstr& instr, u32& addr)
    return false;
 }

-bool DecodeBranch(bool thumb, const FetchedInstr& instr, u32& cond, bool hasLink, u32 lr, bool& link,
+bool DecodeBranch(bool thumb, const FetchedInstr& instr, u32& cond, bool hasLink, u32 lr, bool& link, 
    u32& linkAddr, u32& targetAddr)
 {
    if (thumb)
@ -332,7 +399,7 @@ bool DecodeBranch(bool thumb, const FetchedInstr& instr, u32& cond, bool hasLink
        linkAddr = instr.Addr + 4;

        cond = instr.Cond();
-        if (instr.Info.Kind == ARMInstrInfo::ak_BL
+        if (instr.Info.Kind == ARMInstrInfo::ak_BL 
            || instr.Info.Kind == ARMInstrInfo::ak_B)
        {
            s32 offset = (s32)(instr.Instr << 8) >> 6;
@ -373,7 +440,7 @@ bool IsIdleLoop(bool thumb, FetchedInstr* instrs, int instrsCount)
        u16 dstRegs = instrs[i].Info.DstRegs & ~(1 << 15);

        regsDisallowedToWrite |= srcRegs & ~regsWrittenTo;
-
+        
        if (dstRegs & regsDisallowedToWrite)
            return false;
        regsWrittenTo |= dstRegs;
@ -463,22 +530,12 @@ InterpreterFunc InterpretTHUMB[ARMInstrInfo::tk_Count] =
    F(LDRB_IMM), F(STRH_IMM), F(LDRH_IMM), F(STR_SPREL), F(LDR_SPREL),
    F(PUSH), F(POP), F(LDMIA), F(STMIA),
    F(BCOND), F(BX), F(BLX_REG), F(B), F(BL_LONG_1), F(BL_LONG_2),
-    F(UNK), F(SVC),
+    F(UNK), F(SVC), 
    T_BL_LONG // BL_LONG psudo opcode
 };
 #undef F

-ARMJIT::ARMJIT(melonDS::NDS& nds, std::optional<JITArgs> jit) noexcept : 
-        NDS(nds),
-        Memory(nds),
-        JITCompiler(nds),
-        MaxBlockSize(jit.has_value() ? std::clamp(jit->MaxBlockSize, 1u, 32u) : 32),
-        LiteralOptimizations(jit.has_value() ? jit->LiteralOptimizations : false),
-        BranchOptimizations(jit.has_value() ? jit->BranchOptimizations : false),
-        FastMemory((jit.has_value() ? jit->FastMemory : false) && ARMJIT_Memory::IsFastMemSupported())
-{}
-
-void ARMJIT::RetireJitBlock(JitBlock* block) noexcept
+void RetireJitBlock(JitBlock* block)
 {
    auto it = RestoreCandidates.find(block->InstrHash);
    if (it != RestoreCandidates.end())
@ -492,53 +549,21 @@ void ARMJIT::RetireJitBlock(JitBlock* block) noexcept
    }
 }

-void ARMJIT::SetJITArgs(JITArgs args) noexcept
-{
-    args.FastMemory = args.FastMemory && ARMJIT_Memory::IsFastMemSupported();
-    args.MaxBlockSize = std::clamp(args.MaxBlockSize, 1u, 32u);
-
-    if (MaxBlockSize != args.MaxBlockSize
-        || LiteralOptimizations != args.LiteralOptimizations
-        || BranchOptimizations != args.BranchOptimizations
-        || FastMemory != args.FastMemory)
-        ResetBlockCache();
-
-    MaxBlockSize = args.MaxBlockSize;
-    LiteralOptimizations = args.LiteralOptimizations;
-    BranchOptimizations = args.BranchOptimizations;
-    FastMemory = args.FastMemory;
-}
-
-void ARMJIT::SetMaxBlockSize(int size) noexcept
-{
-    SetJITArgs(JITArgs{static_cast<unsigned>(size), LiteralOptimizations, LiteralOptimizations, FastMemory});
-}
-
-void ARMJIT::SetLiteralOptimizations(bool enabled) noexcept
-{
-    SetJITArgs(JITArgs{static_cast<unsigned>(MaxBlockSize), enabled, BranchOptimizations, FastMemory});
-}
-
-void ARMJIT::SetBranchOptimizations(bool enabled) noexcept
-{
-    SetJITArgs(JITArgs{static_cast<unsigned>(MaxBlockSize), LiteralOptimizations, enabled, FastMemory});
-}
-
-void ARMJIT::SetFastMemory(bool enabled) noexcept
-{
-    SetJITArgs(JITArgs{static_cast<unsigned>(MaxBlockSize), LiteralOptimizations, BranchOptimizations, enabled});
-}
-
-void ARMJIT::CompileBlock(ARM* cpu) noexcept
+void CompileBlock(ARM* cpu)
 {
    bool thumb = cpu->CPSR & 0x20;

+    if (Config::JIT_MaxBlockSize < 1)
+        Config::JIT_MaxBlockSize = 1;
+    if (Config::JIT_MaxBlockSize > 32)
+        Config::JIT_MaxBlockSize = 32;
+
    u32 blockAddr = cpu->R[15] - (thumb ? 2 : 4);

    u32 localAddr = LocaliseCodeAddress(cpu->Num, blockAddr);
    if (!localAddr)
    {
-        Log(LogLevel::Warn, "trying to compile non executable code? %x\n", blockAddr);
+        printf("trying to compile non executable code? %x\n", blockAddr);
    }

    auto& map = cpu->Num == 0 ? JitBlocks9 : JitBlocks7;
@ -556,34 +581,29 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept

            u64* entry = &FastBlockLookupRegions[localAddr >> 27][(localAddr & 0x7FFFFFF) / 2];
            *entry = ((u64)blockAddr | cpu->Num) << 32;
-            *entry |= JITCompiler.SubEntryOffset(existingBlockIt->second->EntryPoint);
+            *entry |= JITCompiler->SubEntryOffset(existingBlockIt->second->EntryPoint);
            return;
        }

        // some memory has been remapped
-        RetireJitBlock(existingBlockIt->second);
+        RetireJitBlock(existingBlockIt->second);        
        map.erase(existingBlockIt);
    }

-    FetchedInstr instrs[MaxBlockSize];
+    FetchedInstr instrs[Config::JIT_MaxBlockSize];
    int i = 0;
    u32 r15 = cpu->R[15];

-    u32 addressRanges[MaxBlockSize];
-    u32 addressMasks[MaxBlockSize];
-    memset(addressMasks, 0, MaxBlockSize * sizeof(u32));
+    u32 addressRanges[Config::JIT_MaxBlockSize];
+    u32 addressMasks[Config::JIT_MaxBlockSize];
+    memset(addressMasks, 0, Config::JIT_MaxBlockSize * sizeof(u32));
    u32 numAddressRanges = 0;

    u32 numLiterals = 0;
-    u32 literalLoadAddrs[MaxBlockSize];
+    u32 literalLoadAddrs[Config::JIT_MaxBlockSize];
    // they are going to be hashed
-    u32 literalValues[MaxBlockSize];
-    u32 instrValues[MaxBlockSize];
-    // due to instruction merging i might not reflect the amount of actual instructions
-    u32 numInstrs = 0;
-
-    u32 writeAddrs[MaxBlockSize];
-    u32 numWriteAddrs = 0, writeAddrsTranslated = 0;
+    u32 literalValues[Config::JIT_MaxBlockSize];
+    u32 instrValues[Config::JIT_MaxBlockSize];

    cpu->FillPipeline();
    u32 nextInstr[2] = {cpu->NextInstr[0], cpu->NextInstr[1]};
@ -595,8 +615,6 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
    u32 lr;
    bool hasLink = false;

-    bool hasMemoryInstr = false;
-
    do
    {
        r15 += thumb ? 2 : 4;
@ -605,13 +623,13 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
        instrs[i].SetFlags = 0;
        instrs[i].Instr = nextInstr[0];
        nextInstr[0] = nextInstr[1];
-
+    
        instrs[i].Addr = nextInstrAddr[0];
        nextInstrAddr[0] = nextInstrAddr[1];
        nextInstrAddr[1] = r15;
        JIT_DEBUGPRINT("instr %08x %x\n", instrs[i].Instr & (thumb ? 0xFFFF : ~0), instrs[i].Addr);

-        instrValues[numInstrs++] = instrs[i].Instr;
+        instrValues[i] = instrs[i].Instr;

        u32 translatedAddr = LocaliseCodeAddress(cpu->Num, instrs[i].Addr);
        assert(translatedAddr >> 27);
@ -619,7 +637,7 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
        if (i == 0 || translatedAddrRounded != addressRanges[numAddressRanges - 1])
        {
            bool returning = false;
-            for (u32 j = 0; j < numAddressRanges; j++)
+            for (int j = 0; j < numAddressRanges; j++)
            {
                if (addressRanges[j] == translatedAddrRounded)
                {
@ -657,11 +675,7 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
                nextInstr[1] = cpuv4->CodeRead32(r15);
            instrs[i].CodeCycles = cpu->CodeCycles;
        }
-        instrs[i].Info = ARMInstrInfo::Decode(thumb, cpu->Num, instrs[i].Instr, LiteralOptimizations);
-
-        hasMemoryInstr |= thumb
-            ? (instrs[i].Info.Kind >= ARMInstrInfo::tk_LDR_PCREL && instrs[i].Info.Kind <= ARMInstrInfo::tk_STMIA)
-            : (instrs[i].Info.Kind >= ARMInstrInfo::ak_STR_REG_LSL && instrs[i].Info.Kind <= ARMInstrInfo::ak_STM);
+        instrs[i].Info = ARMInstrInfo::Decode(thumb, cpu->Num, instrs[i].Instr);

        cpu->R[15] = r15;
        cpu->CurInstr = instrs[i].Instr;
@ -701,47 +715,41 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
        instrs[i].DataRegion = cpu->DataRegion;

        u32 literalAddr;
-        if (LiteralOptimizations
+        if (Config::JIT_LiteralOptimisations
            && instrs[i].Info.SpecialKind == ARMInstrInfo::special_LoadLiteral
            && DecodeLiteral(thumb, instrs[i], literalAddr))
        {
            u32 translatedAddr = LocaliseCodeAddress(cpu->Num, literalAddr);
            if (!translatedAddr)
            {
-                Log(LogLevel::Warn,"literal in non executable memory?\n");
+                printf("literal in non executable memory?\n");
            }
-            if (InvalidLiterals.Find(translatedAddr) == -1)
-            {
-                u32 translatedAddrRounded = translatedAddr & ~0x1FF;
+            u32 translatedAddrRounded = translatedAddr & ~0x1FF;

-                u32 j = 0;
-                for (; j < numAddressRanges; j++)
-                    if (addressRanges[j] == translatedAddrRounded)
-                        break;
-                if (j == numAddressRanges)
-                    addressRanges[numAddressRanges++] = translatedAddrRounded;
-                addressMasks[j] |= 1 << ((translatedAddr & 0x1FF) / 16);
-                JIT_DEBUGPRINT("literal loading %08x %08x %08x %08x\n", literalAddr, translatedAddr, addressMasks[j], addressRanges[j]);
-                cpu->DataRead32(literalAddr, &literalValues[numLiterals]);
-                literalLoadAddrs[numLiterals++] = translatedAddr;
-            }
+            u32 j = 0;
+            for (; j < numAddressRanges; j++)
+                if (addressRanges[j] == translatedAddrRounded)
+                    break;
+            if (j == numAddressRanges)
+                addressRanges[numAddressRanges++] = translatedAddrRounded;
+            addressMasks[j] |= 1 << ((translatedAddr & 0x1FF) / 16);
+            JIT_DEBUGPRINT("literal loading %08x %08x %08x %08x\n", literalAddr, translatedAddr, addressMasks[j], addressRanges[j]);
+            cpu->DataRead32(literalAddr, &literalValues[numLiterals]);
+            literalLoadAddrs[numLiterals++] = translatedAddr;
        }
-        else if (instrs[i].Info.SpecialKind == ARMInstrInfo::special_WriteMem)
-            writeAddrs[numWriteAddrs++] = instrs[i].DataRegion;
-        else if (thumb && instrs[i].Info.Kind == ARMInstrInfo::tk_BL_LONG_2 && i > 0
+
+        if (thumb && instrs[i].Info.Kind == ARMInstrInfo::tk_BL_LONG_2 && i > 0
            && instrs[i - 1].Info.Kind == ARMInstrInfo::tk_BL_LONG_1)
        {
+            instrs[i - 1].Info.Kind = ARMInstrInfo::tk_BL_LONG;
+            instrs[i - 1].Instr = (instrs[i - 1].Instr & 0xFFFF) | (instrs[i].Instr << 16);
+            instrs[i - 1].Info.DstRegs = 0xC000;
+            instrs[i - 1].Info.SrcRegs = 0;
+            instrs[i - 1].Info.EndBlock = true;
            i--;
-            instrs[i].Info.Kind = ARMInstrInfo::tk_BL_LONG;
-            instrs[i].Instr = (instrs[i].Instr & 0xFFFF) | (instrs[i + 1].Instr << 16);
-            instrs[i].Info.DstRegs = 0xC000;
-            instrs[i].Info.SrcRegs = 0;
-            instrs[i].Info.EndBlock = true;
-            JIT_DEBUGPRINT("merged BL\n");
        }

-        if (instrs[i].Info.Branches() && BranchOptimizations
-            && instrs[i].Info.Kind != (thumb ? ARMInstrInfo::tk_SVC : ARMInstrInfo::ak_SVC))
+        if (instrs[i].Info.Branches() && Config::JIT_BranchOptimisations)
        {
            bool hasBranched = cpu->R[15] != r15;

@ -759,7 +767,7 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
                {
                    for (int j = 0; j < i; j++)
                    {
-                        if (instrs[j].Addr == target)
+                        if (instrs[i].Addr == target)
                        {
                            isBackJump = true;
                            break;
@ -777,14 +785,14 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
                        JIT_DEBUGPRINT("found %s idle loop %d in block %08x\n", thumb ? "thumb" : "arm", cpu->Num, blockAddr);
                    }
                }
-                else if (hasBranched && !isBackJump && i + 1 < MaxBlockSize)
+                else if (hasBranched && !isBackJump && i + 1 < Config::JIT_MaxBlockSize)
                {
                    if (link)
                    {
                        lr = linkAddr;
                        hasLink = true;
                    }
-
+                    
                    r15 = target + (thumb ? 2 : 4);
                    assert(r15 == cpu->R[15]);

@ -805,9 +813,8 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
                }
            }

-            if (!hasBranched && cond < 0xE && i + 1 < MaxBlockSize)
+            if (!hasBranched && cond < 0xE && i + 1 < Config::JIT_MaxBlockSize)
            {
-                JIT_DEBUGPRINT("block lengthened by untaken branch\n");
                instrs[i].Info.EndBlock = false;
                instrs[i].BranchFlags |= branch_FollowCondNotTaken;
            }
@ -815,34 +822,14 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept

        i++;

-        bool canCompile = JITCompiler.CanCompile(thumb, instrs[i - 1].Info.Kind);
+        bool canCompile = JITCompiler->CanCompile(thumb, instrs[i - 1].Info.Kind);
        bool secondaryFlagReadCond = !canCompile || (instrs[i - 1].BranchFlags & (branch_FollowCondTaken | branch_FollowCondNotTaken));
        if (instrs[i - 1].Info.ReadFlags != 0 || secondaryFlagReadCond)
            FloodFillSetFlags(instrs, i - 2, !secondaryFlagReadCond ? instrs[i - 1].Info.ReadFlags : 0xF);
-    } while(!instrs[i - 1].Info.EndBlock && i < MaxBlockSize && !cpu->Halted && (!cpu->IRQ || (cpu->CPSR & 0x80)));
-
-    if (numLiterals)
-    {
-        for (u32 j = 0; j < numWriteAddrs; j++)
-        {
-            u32 translatedAddr = LocaliseCodeAddress(cpu->Num, writeAddrs[j]);
-            if (translatedAddr)
-            {
-                for (u32 k = 0; k < numLiterals; k++)
-                {
-                    if (literalLoadAddrs[k] == translatedAddr)
-                    {
-                        if (InvalidLiterals.Find(translatedAddr) == -1)
-                            InvalidLiterals.Add(translatedAddr);
-                        break;
-                    }
-                }
-            }
-        }
-    }
+    } while(!instrs[i - 1].Info.EndBlock && i < Config::JIT_MaxBlockSize && !cpu->Halted && (!cpu->IRQ || (cpu->CPSR & 0x80)));

    u32 literalHash = (u32)XXH3_64bits(literalValues, numLiterals * 4);
-    u32 instrHash = (u32)XXH3_64bits(instrValues, numInstrs * 4);
+    u32 instrHash = (u32)XXH3_64bits(instrValues, i * 4);

    auto prevBlockIt = RestoreCandidates.find(instrHash);
    JitBlock* prevBlock = NULL;
@ -856,7 +843,7 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept

        if (mayRestore && prevBlock->NumAddresses == numAddressRanges)
        {
-            for (u32 j = 0; j < numAddressRanges; j++)
+            for (int j = 0; j < numAddressRanges; j++)
            {
                if (prevBlock->AddressRanges()[j] != addressRanges[j]
                    || prevBlock->AddressMasks()[j] != addressMasks[j])
@ -883,9 +870,9 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
        block = new JitBlock(cpu->Num, i, numAddressRanges, numLiterals);
        block->LiteralHash = literalHash;
        block->InstrHash = instrHash;
-        for (u32 j = 0; j < numAddressRanges; j++)
+        for (int j = 0; j < numAddressRanges; j++)
            block->AddressRanges()[j] = addressRanges[j];
-        for (u32 j = 0; j < numAddressRanges; j++)
+        for (int j = 0; j < numAddressRanges; j++)
            block->AddressMasks()[j] = addressMasks[j];
        for (int j = 0; j < numLiterals; j++)
            block->Literals()[j] = literalLoadAddrs[j];
@ -895,9 +882,7 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept

        FloodFillSetFlags(instrs, i - 1, 0xF);

-        JitEnableWrite();
-        block->EntryPoint = JITCompiler.CompileBlock(cpu, thumb, instrs, i, hasMemoryInstr);
-        JitEnableExecute();
+        block->EntryPoint = JITCompiler->CompileBlock(cpu, thumb, instrs, i);

        JIT_DEBUGPRINT("block start %p\n", block->EntryPoint);
    }
@ -908,7 +893,7 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept
    }

    assert((localAddr & 1) == 0);
-    for (u32 j = 0; j < numAddressRanges; j++)
+    for (int j = 0; j < numAddressRanges; j++)
    {
        assert(addressRanges[j] == block->AddressRanges()[j]);
        assert(addressMasks[j] == block->AddressMasks()[j]);
@ -916,8 +901,8 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept

        AddressRange* region = CodeMemRegions[addressRanges[j] >> 27];

-        if (!PageContainsCode(&region[(addressRanges[j] & 0x7FFF000 & ~(Memory.PageSize - 1)) / 512], Memory.PageSize))
-            Memory.SetCodeProtection(addressRanges[j] >> 27, addressRanges[j] & 0x7FFFFFF, true);
+        if (!PageContainsCode(&region[(addressRanges[j] & 0x7FFF000) / 512]))
+            ARMJIT_Memory::SetCodeProtection(addressRanges[j] >> 27, addressRanges[j] & 0x7FFFFFF, true);

        AddressRange* range = &region[(addressRanges[j] & 0x7FFFFFF) / 512];
        range->Code |= addressMasks[j];
@ -931,10 +916,10 @@ void ARMJIT::CompileBlock(ARM* cpu) noexcept

    u64* entry = &FastBlockLookupRegions[(localAddr >> 27)][(localAddr & 0x7FFFFFF) / 2];
    *entry = ((u64)blockAddr | cpu->Num) << 32;
-    *entry |= JITCompiler.SubEntryOffset(block->EntryPoint);
+    *entry |= JITCompiler->SubEntryOffset(block->EntryPoint);
 }

-void ARMJIT::InvalidateByAddr(u32 localAddr) noexcept
+void InvalidateByAddr(u32 localAddr)
 {
    JIT_DEBUGPRINT("invalidating by addr %x\n", localAddr);

@ -969,9 +954,9 @@ void ARMJIT::InvalidateByAddr(u32 localAddr) noexcept
        range->Blocks.Remove(i);

        if (range->Blocks.Length == 0
-            && !PageContainsCode(&region[(localAddr & 0x7FFF000 & ~(Memory.PageSize - 1)) / 512], Memory.PageSize))
+            && !PageContainsCode(&region[(localAddr & 0x7FFF000) / 512]))
        {
-            Memory.SetCodeProtection(localAddr >> 27, localAddr & 0x7FFFFFF, false);
+            ARMJIT_Memory::SetCodeProtection(localAddr >> 27, localAddr & 0x7FFFFFF, false);
        }

        bool literalInvalidation = false;
@ -980,7 +965,7 @@ void ARMJIT::InvalidateByAddr(u32 localAddr) noexcept
            u32 addr = block->Literals()[j];
            if (addr == localAddr)
            {
-                if (InvalidLiterals.Find(localAddr) == -1)
+                if (InvalidLiterals.Find(localAddr) != -1)
                {
                    InvalidLiterals.Add(localAddr);
                    JIT_DEBUGPRINT("found invalid literal %d\n", InvalidLiterals.Length);
@ -1003,8 +988,8 @@ void ARMJIT::InvalidateByAddr(u32 localAddr) noexcept

                if (otherRange->Blocks.Length == 0)
                {
-                    if (!PageContainsCode(&otherRegion[(addr & 0x7FFF000 & ~(Memory.PageSize - 1)) / 512], Memory.PageSize))
-                        Memory.SetCodeProtection(addr >> 27, addr & 0x7FFFFFF, false);
+                    if (!PageContainsCode(&otherRegion[(addr & 0x7FFF000) / 512]))
+                        ARMJIT_Memory::SetCodeProtection(addr >> 27, addr & 0x7FFFFFF, false);

                    otherRange->Code = 0;
                }
@ -1028,64 +1013,49 @@ void ARMJIT::InvalidateByAddr(u32 localAddr) noexcept
    }
 }

-void ARMJIT::CheckAndInvalidateITCM() noexcept
+void CheckAndInvalidateITCM()
 {
-    for (u32 i = 0; i < ITCMPhysicalSize; i+=512)
+    for (u32 i = 0; i < ITCMPhysicalSize; i+=16)
    {
-        if (CodeIndexITCM[i / 512].Code)
+        if (CodeIndexITCM[i / 512].Code & (1 << ((i & 0x1FF) / 16)))
        {
-            // maybe using bitscan would be better here?
-            // The thing is that in densely populated sets
-            // The old fashioned way can actually be faster
-            for (u32 j = 0; j < 512; j += 16)
-            {
-                if (CodeIndexITCM[i / 512].Code & (1 << ((j & 0x1FF) / 16)))
-                    InvalidateByAddr((i+j) | (ARMJIT_Memory::memregion_ITCM << 27));
-            }
+            InvalidateByAddr(i | (ARMJIT_Memory::memregion_ITCM << 27));
        }
    }
 }

-void ARMJIT::CheckAndInvalidateWVRAM(int bank) noexcept
+template <u32 num, int region>
+void CheckAndInvalidate(u32 addr)
 {
-    u32 start = bank == 1 ? 0x20000 : 0;
-    for (u32 i = start; i < start+0x20000; i+=512)
-    {
-        if (CodeIndexARM7WVRAM[i / 512].Code)
-        {
-            for (u32 j = 0; j < 512; j += 16)
-            {
-                if (CodeIndexARM7WVRAM[i / 512].Code & (1 << ((j & 0x1FF) / 16)))
-                    InvalidateByAddr((i+j) | (ARMJIT_Memory::memregion_VWRAM << 27));
-            }
-        }
-    }
+    u32 localAddr = ARMJIT_Memory::LocaliseAddress(region, num, addr);
+    if (CodeMemRegions[region][(localAddr & 0x7FFFFFF) / 512].Code & (1 << ((localAddr & 0x1FF) / 16)))
+        InvalidateByAddr(localAddr);
 }

-JitBlockEntry ARMJIT::LookUpBlock(u32 num, u64* entries, u32 offset, u32 addr) noexcept
+JitBlockEntry LookUpBlock(u32 num, u64* entries, u32 offset, u32 addr)
 {
    u64* entry = &entries[offset / 2];
    if (*entry >> 32 == (addr | num))
-        return JITCompiler.AddEntryOffset((u32)*entry);
+        return JITCompiler->AddEntryOffset((u32)*entry);
    return NULL;
 }

-void ARMJIT::blockSanityCheck(u32 num, u32 blockAddr, JitBlockEntry entry) noexcept
+void blockSanityCheck(u32 num, u32 blockAddr, JitBlockEntry entry)
 {
    u32 localAddr = LocaliseCodeAddress(num, blockAddr);
-    assert(JITCompiler.AddEntryOffset((u32)FastBlockLookupRegions[localAddr >> 27][(localAddr & 0x7FFFFFF) / 2]) == entry);
+    assert(JITCompiler->AddEntryOffset((u32)FastBlockLookupRegions[localAddr >> 27][(localAddr & 0x7FFFFFF) / 2]) == entry);
 }

-bool ARMJIT::SetupExecutableRegion(u32 num, u32 blockAddr, u64*& entry, u32& start, u32& size) noexcept
+bool SetupExecutableRegion(u32 num, u32 blockAddr, u64*& entry, u32& start, u32& size)
 {
    // amazingly ignoring the DTCM is the proper behaviour for code fetches
    int region = num == 0
-        ? Memory.ClassifyAddress9(blockAddr)
-        : Memory.ClassifyAddress7(blockAddr);
+        ? ARMJIT_Memory::ClassifyAddress9(blockAddr)
+        : ARMJIT_Memory::ClassifyAddress7(blockAddr);

    u32 memoryOffset;
    if (FastBlockLookupRegions[region]
-        && Memory.GetMirrorLocation(region, num, blockAddr, memoryOffset, start, size))
+        && ARMJIT_Memory::GetMirrorLocation(region, num, blockAddr, memoryOffset, start, size))
    {
        //printf("setup exec region %d %d %08x %08x %x %x\n", num, region, blockAddr, start, size, memoryOffset);
        entry = FastBlockLookupRegions[region] + memoryOffset / 2;
@ -1094,28 +1064,28 @@ bool ARMJIT::SetupExecutableRegion(u32 num, u32 blockAddr, u64*& entry, u32& sta
    return false;
 }

-template void ARMJIT::CheckAndInvalidate<0, ARMJIT_Memory::memregion_MainRAM>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<1, ARMJIT_Memory::memregion_MainRAM>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<0, ARMJIT_Memory::memregion_SharedWRAM>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<1, ARMJIT_Memory::memregion_SharedWRAM>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<1, ARMJIT_Memory::memregion_WRAM7>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<1, ARMJIT_Memory::memregion_VWRAM>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<0, ARMJIT_Memory::memregion_VRAM>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<0, ARMJIT_Memory::memregion_ITCM>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<0, ARMJIT_Memory::memregion_NewSharedWRAM_A>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<1, ARMJIT_Memory::memregion_NewSharedWRAM_A>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<0, ARMJIT_Memory::memregion_NewSharedWRAM_B>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<1, ARMJIT_Memory::memregion_NewSharedWRAM_B>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<0, ARMJIT_Memory::memregion_NewSharedWRAM_C>(u32) noexcept;
-template void ARMJIT::CheckAndInvalidate<1, ARMJIT_Memory::memregion_NewSharedWRAM_C>(u32) noexcept;
+template void CheckAndInvalidate<0, ARMJIT_Memory::memregion_MainRAM>(u32);
+template void CheckAndInvalidate<1, ARMJIT_Memory::memregion_MainRAM>(u32);
+template void CheckAndInvalidate<0, ARMJIT_Memory::memregion_SharedWRAM>(u32);
+template void CheckAndInvalidate<1, ARMJIT_Memory::memregion_SharedWRAM>(u32);
+template void CheckAndInvalidate<1, ARMJIT_Memory::memregion_WRAM7>(u32);
+template void CheckAndInvalidate<1, ARMJIT_Memory::memregion_VWRAM>(u32);
+template void CheckAndInvalidate<0, ARMJIT_Memory::memregion_VRAM>(u32);
+template void CheckAndInvalidate<0, ARMJIT_Memory::memregion_ITCM>(u32);
+template void CheckAndInvalidate<0, ARMJIT_Memory::memregion_NewSharedWRAM_A>(u32);
+template void CheckAndInvalidate<1, ARMJIT_Memory::memregion_NewSharedWRAM_A>(u32);
+template void CheckAndInvalidate<0, ARMJIT_Memory::memregion_NewSharedWRAM_B>(u32);
+template void CheckAndInvalidate<1, ARMJIT_Memory::memregion_NewSharedWRAM_B>(u32);
+template void CheckAndInvalidate<0, ARMJIT_Memory::memregion_NewSharedWRAM_C>(u32);
+template void CheckAndInvalidate<1, ARMJIT_Memory::memregion_NewSharedWRAM_C>(u32);

-void ARMJIT::ResetBlockCache() noexcept
+void ResetBlockCache()
 {
-    Log(LogLevel::Debug, "Resetting JIT block cache...\n");
+    printf("Resetting JIT block cache...\n");

    // could be replace through a function which only resets
    // the permissions but we're too lazy
-    Memory.Reset();
+    ARMJIT_Memory::Reset();

    InvalidLiterals.Clear();
    for (int i = 0; i < ARMJIT_Memory::memregions_Count; i++)
@ -1153,23 +1123,7 @@ void ARMJIT::ResetBlockCache() noexcept
    JitBlocks9.clear();
    JitBlocks7.clear();

-    JITCompiler.Reset();
-}
-
-void ARMJIT::JitEnableWrite() noexcept
-{
-    #if defined(__APPLE__) && defined(__aarch64__)
-        if (__builtin_available(macOS 11.0, *))
-            pthread_jit_write_protect_np(false);
-    #endif
-}
-
-void ARMJIT::JitEnableExecute() noexcept
-{
-    #if defined(__APPLE__) && defined(__aarch64__)
-        if (__builtin_available(macOS 11.0, *))
-            pthread_jit_write_protect_np(true);
-    #endif
+    JITCompiler->Reset();
 }

 }
--- a/src/ARMJIT.h
+++ b/src/ARMJIT.h
@ -1,206 +1,37 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #ifndef ARMJIT_H
 #define ARMJIT_H

-#include <algorithm>
-#include <optional>
-#include <memory>
 #include "types.h"
-#include "MemConstants.h"
-#include "Args.h"
-#include "ARMJIT_Memory.h"

-#ifdef JIT_ENABLED
-#include "JitBlock.h"
+#include "ARM.h"
+#include "ARM_InstrInfo.h"

-#if defined(__APPLE__) && defined(__aarch64__)
-    #include <pthread.h>
-#endif
-
-#include "ARMJIT_Compiler.h"
-
-namespace melonDS
+namespace ARMJIT
 {
-class ARM;

-class JitBlock;
-class ARMJIT
-{
-public:
-    ARMJIT(melonDS::NDS& nds, std::optional<JITArgs> jit) noexcept;
-    ~ARMJIT() noexcept;
-    void InvalidateByAddr(u32) noexcept;
-    void CheckAndInvalidateWVRAM(int) noexcept;
-    void CheckAndInvalidateITCM() noexcept;
-    void Reset() noexcept;
-    void JitEnableWrite() noexcept;
-    void JitEnableExecute() noexcept;
-    void CompileBlock(ARM* cpu) noexcept;
-    void ResetBlockCache() noexcept;
+typedef void (*JitBlockEntry)();

-    template <u32 num, int region>
-    void CheckAndInvalidate(u32 addr) noexcept
-    {
-        u32 localAddr = Memory.LocaliseAddress(region, num, addr);
-        if (CodeMemRegions[region][(localAddr & 0x7FFFFFF) / 512].Code & (1 << ((localAddr & 0x1FF) / 16)))
-            InvalidateByAddr(localAddr);
-    }
-    JitBlockEntry LookUpBlock(u32 num, u64* entries, u32 offset, u32 addr) noexcept;
-    bool SetupExecutableRegion(u32 num, u32 blockAddr, u64*& entry, u32& start, u32& size) noexcept;
-    u32 LocaliseCodeAddress(u32 num, u32 addr) const noexcept;
+void Init();
+void DeInit();

-    ARMJIT_Memory Memory;
-private:
-    int MaxBlockSize {};
-    bool LiteralOptimizations = false;
-    bool BranchOptimizations = false;
-    bool FastMemory = false;
+void Reset();

-public:
-    melonDS::NDS& NDS;
-    TinyVector<u32> InvalidLiterals {};
-    friend class ARMJIT_Memory;
-    void blockSanityCheck(u32 num, u32 blockAddr, JitBlockEntry entry) noexcept;
-    void RetireJitBlock(JitBlock* block) noexcept;
+void CheckAndInvalidateITCM();

-    int GetMaxBlockSize() const noexcept { return MaxBlockSize; }
-    bool LiteralOptimizationsEnabled() const noexcept { return LiteralOptimizations; }
-    bool BranchOptimizationsEnabled() const noexcept { return BranchOptimizations; }
-    bool FastMemoryEnabled() const noexcept { return FastMemory; }
+void InvalidateByAddr(u32 pseudoPhysical);

-    void SetJITArgs(JITArgs args) noexcept;
-    void SetMaxBlockSize(int size) noexcept;
-    void SetLiteralOptimizations(bool enabled) noexcept;
-    void SetBranchOptimizations(bool enabled) noexcept;
-    void SetFastMemory(bool enabled) noexcept;
+template <u32 num, int region>
+void CheckAndInvalidate(u32 addr);

-    Compiler JITCompiler;
-    std::unordered_map<u32, JitBlock*> JitBlocks9 {};
-    std::unordered_map<u32, JitBlock*> JitBlocks7 {};
+void CompileBlock(ARM* cpu);

-    std::unordered_map<u32, JitBlock*> RestoreCandidates {};
+void ResetBlockCache();

+JitBlockEntry LookUpBlock(u32 num, u64* entries, u32 offset, u32 addr);
+bool SetupExecutableRegion(u32 num, u32 blockAddr, u64*& entry, u32& start, u32& size);

-    AddressRange CodeIndexITCM[ITCMPhysicalSize / 512] {};
-    AddressRange CodeIndexMainRAM[MainRAMMaxSize / 512] {};
-    AddressRange CodeIndexSWRAM[SharedWRAMSize / 512] {};
-    AddressRange CodeIndexVRAM[0x100000 / 512] {};
-    AddressRange CodeIndexARM9BIOS[ARM9BIOSSize / 512] {};
-    AddressRange CodeIndexARM7BIOS[ARM7BIOSSize / 512] {};
-    AddressRange CodeIndexARM7WRAM[ARM7WRAMSize / 512] {};
-    AddressRange CodeIndexARM7WVRAM[0x40000 / 512] {};
-    AddressRange CodeIndexBIOS9DSi[0x10000 / 512] {};
-    AddressRange CodeIndexBIOS7DSi[0x10000 / 512] {};
-    AddressRange CodeIndexNWRAM_A[NWRAMSize / 512] {};
-    AddressRange CodeIndexNWRAM_B[NWRAMSize / 512] {};
-    AddressRange CodeIndexNWRAM_C[NWRAMSize / 512] {};
-
-    u64 FastBlockLookupITCM[ITCMPhysicalSize / 2] {};
-    u64 FastBlockLookupMainRAM[MainRAMMaxSize / 2] {};
-    u64 FastBlockLookupSWRAM[SharedWRAMSize / 2] {};
-    u64 FastBlockLookupVRAM[0x100000 / 2] {};
-    u64 FastBlockLookupARM9BIOS[ARM9BIOSSize / 2] {};
-    u64 FastBlockLookupARM7BIOS[ARM7BIOSSize / 2] {};
-    u64 FastBlockLookupARM7WRAM[ARM7WRAMSize / 2] {};
-    u64 FastBlockLookupARM7WVRAM[0x40000 / 2] {};
-    u64 FastBlockLookupBIOS9DSi[0x10000 / 2] {};
-    u64 FastBlockLookupBIOS7DSi[0x10000 / 2] {};
-    u64 FastBlockLookupNWRAM_A[NWRAMSize / 2] {};
-    u64 FastBlockLookupNWRAM_B[NWRAMSize / 2] {};
-    u64 FastBlockLookupNWRAM_C[NWRAMSize / 2] {};
-
-    AddressRange* const CodeMemRegions[ARMJIT_Memory::memregions_Count] =
-    {
-        NULL,
-        CodeIndexITCM,
-        NULL,
-        CodeIndexARM9BIOS,
-        CodeIndexMainRAM,
-        CodeIndexSWRAM,
-        NULL,
-        CodeIndexVRAM,
-        CodeIndexARM7BIOS,
-        CodeIndexARM7WRAM,
-        NULL,
-        NULL,
-        CodeIndexARM7WVRAM,
-        CodeIndexBIOS9DSi,
-        CodeIndexBIOS7DSi,
-        CodeIndexNWRAM_A,
-        CodeIndexNWRAM_B,
-        CodeIndexNWRAM_C
-    };
-
-    u64* const FastBlockLookupRegions[ARMJIT_Memory::memregions_Count] =
-    {
-        NULL,
-        FastBlockLookupITCM,
-        NULL,
-        FastBlockLookupARM9BIOS,
-        FastBlockLookupMainRAM,
-        FastBlockLookupSWRAM,
-        NULL,
-        FastBlockLookupVRAM,
-        FastBlockLookupARM7BIOS,
-        FastBlockLookupARM7WRAM,
-        NULL,
-        NULL,
-        FastBlockLookupARM7WVRAM,
-        FastBlockLookupBIOS9DSi,
-        FastBlockLookupBIOS7DSi,
-        FastBlockLookupNWRAM_A,
-        FastBlockLookupNWRAM_B,
-        FastBlockLookupNWRAM_C
-    };
-};
 }

-// Defined in assembly
-extern "C" void ARM_Dispatch(melonDS::ARM* cpu, melonDS::JitBlockEntry entry);
-#else
-namespace melonDS
-{
-class ARM;
-
-// This version is a stub; the methods all do nothing,
-// but there's still a Memory member.
-class ARMJIT
-{
-public:
-    ARMJIT(melonDS::NDS& nds, std::optional<JITArgs>) noexcept : Memory(nds) {}
-    ~ARMJIT() noexcept {}
-    void InvalidateByAddr(u32) noexcept {}
-    void CheckAndInvalidateWVRAM(int) noexcept {}
-    void CheckAndInvalidateITCM() noexcept {}
-    void Reset() noexcept {}
-    void JitEnableWrite() noexcept {}
-    void JitEnableExecute() noexcept {}
-    void CompileBlock(ARM*) noexcept {}
-    void ResetBlockCache() noexcept {}
-    template <u32, int>
-    void CheckAndInvalidate(u32 addr) noexcept {}
-
-    ARMJIT_Memory Memory;
-};
-}
-#endif // JIT_ENABLED
-
-#endif // ARMJIT_H
+extern "C" void ARM_Dispatch(ARM* cpu, ARMJIT::JitBlockEntry entry);

+#endif
--- a/src/ARMJIT_A64/ARMJIT_ALU.cpp
+++ b/src/ARMJIT_A64/ARMJIT_ALU.cpp
@ -1,26 +1,8 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARMJIT_Compiler.h"

 using namespace Arm64Gen;

-namespace melonDS
+namespace ARMJIT
 {

 void Compiler::Comp_RegShiftReg(int op, bool S, Op2& op2, ARM64Reg rs)
@ -324,29 +306,16 @@ void Compiler::Comp_Arithmetic(int op, bool S, ARM64Reg rd, ARM64Reg rn, Op2 op2
        UBFX(W2, RCPSR, 29, 1);
        if (S)
        {
+            CVInGPR = true;
+            ADDS(W1, rn, W2);
+            CSET(W2, CC_CS);
+            CSET(W3, CC_VS);
            if (op2.IsImm)
-            {
-                CVInGPR = true;
-                ADDS(W1, rn, W2);
-                CSET(W2, CC_CS);
-                CSET(W3, CC_VS);
-                if (op2.IsImm)
-                    ADDSI2R(rd, W1, op2.Imm, W0);
-                else
-                    ADDS(rd, W1, op2.Reg.Rm, op2.ToArithOption());
-                CSINC(W2, W2, WZR, CC_CC);
-                CSINC(W3, W3, WZR, CC_VC);
-            }
+                ADDSI2R(rd, W1, op2.Imm, W0);
            else
-            {
-                if (op2.Reg.ShiftAmount > 0)
-                {
-                    MOV(W0, op2.Reg.Rm, op2.ToArithOption());
-                    op2 = Op2(W0, ST_LSL, 0);
-                }
-                CMP(W2, 1);
-                ADCS(rd, rn, op2.Reg.Rm);
-            }
+                ADDS(rd, W1, op2.Reg.Rm, op2.ToArithOption());
+            CSINC(W2, W2, WZR, CC_CC);
+            CSINC(W3, W3, WZR, CC_VC);
        }
        else
        {
@ -359,38 +328,25 @@ void Compiler::Comp_Arithmetic(int op, bool S, ARM64Reg rd, ARM64Reg rn, Op2 op2
        break;
    case 0x6: // SBC
        UBFX(W2, RCPSR, 29, 1);
-        if (S && !op2.IsImm)
+        // W1 = -op2 - 1
+        if (op2.IsImm)
+            MOVI2R(W1, ~op2.Imm);
+        else
+            ORN(W1, WZR, op2.Reg.Rm, op2.ToArithOption());
+        if (S)
        {
-            if (op2.Reg.ShiftAmount > 0)
-            {
-                MOV(W0, op2.Reg.Rm, op2.ToArithOption());
-                op2 = Op2(W0, ST_LSL, 0);
-            }
-            CMP(W2, 1);
-            SBCS(rd, rn, op2.Reg.Rm);
+            CVInGPR = true;
+            ADDS(W1, W2, W1);
+            CSET(W2, CC_CS);
+            CSET(W3, CC_VS);
+            ADDS(rd, rn, W1);
+            CSINC(W2, W2, WZR, CC_CC);
+            CSINC(W3, W3, WZR, CC_VC);
        }
        else
        {
-            // W1 = -op2 - 1
-            if (op2.IsImm)
-                MOVI2R(W1, ~op2.Imm);
-            else
-                ORN(W1, WZR, op2.Reg.Rm, op2.ToArithOption());
-            if (S)
-            {
-                CVInGPR = true;
-                ADDS(W1, W2, W1);
-                CSET(W2, CC_CS);
-                CSET(W3, CC_VS);
-                ADDS(rd, rn, W1);
-                CSINC(W2, W2, WZR, CC_CC);
-                CSINC(W3, W3, WZR, CC_VC);
-            }
-            else
-            {
-                ADD(W1, W2, W1);
-                ADD(rd, rn, W1);
-            }
+            ADD(W1, W2, W1);
+            ADD(rd, rn, W1);
        }
        break;
    case 0x7: // RSC
@ -480,7 +436,7 @@ void Compiler::A_Comp_GetOp2(bool S, Op2& op2)
        Comp_AddCycles_C();

        u32 shift = (CurInstr.Instr >> 7) & 0x1E;
-        u32 imm = melonDS::ROR(CurInstr.Instr & 0xFF, shift);
+        u32 imm = ::ROR(CurInstr.Instr & 0xFF, shift);

        if (S && shift && (CurInstr.SetFlags & 0x2))
        {
@ -559,7 +515,21 @@ void Compiler::A_Comp_ALUMovOp()
        }
        else
        {
-            MOV(rd, op2.Reg.Rm, op2.ToArithOption());
+            // ORR with shifted operand has cycles latency
+            if (op2.Reg.ShiftAmount > 0)
+            {
+                switch (op2.Reg.ShiftType)
+                {
+                case ST_LSL: LSL(rd, op2.Reg.Rm, op2.Reg.ShiftAmount); break;
+                case ST_LSR: LSR(rd, op2.Reg.Rm, op2.Reg.ShiftAmount); break;
+                case ST_ASR: ASR(rd, op2.Reg.Rm, op2.Reg.ShiftAmount); break;
+                case ST_ROR: ROR(rd, op2.Reg.Rm, op2.Reg.ShiftAmount); break;
+                }
+            }
+            else
+            {
+                MOV(rd, op2.Reg.Rm, op2.ToArithOption());
+            }
        }
    }

--- a/src/ARMJIT_A64/ARMJIT_Branch.cpp
+++ b/src/ARMJIT_A64/ARMJIT_Branch.cpp
@ -1,34 +1,15 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARMJIT_Compiler.h"
-#include "../NDS.h"

 using namespace Arm64Gen;

 // hack
 const int kCodeCacheTiming = 3;

-namespace melonDS
+namespace ARMJIT
 {

 template <typename T>
-void JumpToTrampoline(T* cpu, u32 addr, bool changeCPSR)
+void jumpToTrampoline(T* cpu, u32 addr, bool changeCPSR)
 {
    cpu->JumpTo(addr, changeCPSR);
 }
@ -83,7 +64,7 @@ void Compiler::Comp_JumpTo(u32 addr, bool forceNonConstantCycles)
            // doesn't matter if we put garbage in the MSbs there
            if (addr & 0x2)
            {
-                cpu9->CodeRead32(addr-2, true);
+                cpu9->CodeRead32(addr-2, true) >> 16;
                cycles += cpu9->CodeCycles;
                cpu9->CodeRead32(addr+2, false);
                cycles += CurCPU->CodeCycles;
@ -133,7 +114,7 @@ void Compiler::Comp_JumpTo(u32 addr, bool forceNonConstantCycles)
            u32 compileTimePC = CurCPU->R[15];
            CurCPU->R[15] = newPC;

-            cycles += NDS.ARM7MemTimings[codeCycles][0] + NDS.ARM7MemTimings[codeCycles][1];
+            cycles += NDS::ARM7MemTimings[codeCycles][0] + NDS::ARM7MemTimings[codeCycles][1];

            CurCPU->R[15] = compileTimePC;
        }
@ -145,7 +126,7 @@ void Compiler::Comp_JumpTo(u32 addr, bool forceNonConstantCycles)
            u32 compileTimePC = CurCPU->R[15];
            CurCPU->R[15] = newPC;

-            cycles += NDS.ARM7MemTimings[codeCycles][2] + NDS.ARM7MemTimings[codeCycles][3];
+            cycles += NDS::ARM7MemTimings[codeCycles][2] + NDS::ARM7MemTimings[codeCycles][3];

            CurCPU->R[15] = compileTimePC;
        }
@ -236,7 +217,7 @@ void* Compiler::Gen_JumpTo7(int kind)
    LSR(W1, W0, 15);
    STR(INDEX_UNSIGNED, W1, RCPU, offsetof(ARM, CodeCycles));

-    MOVP2R(X2, NDS.ARM7MemTimings);
+    MOVP2R(X2, NDS::ARM7MemTimings);
    LDR(W3, X2, ArithOption(W1, true));

    FixupBranch switchToThumb;
@ -302,7 +283,7 @@ void Compiler::Comp_JumpTo(Arm64Gen::ARM64Reg addr, bool switchThumb, bool resto
        bool cpsrDirty = CPSRDirty;
        SaveCPSR();
        SaveCycles();
-        PushRegs(restoreCPSR, true);
+        PushRegs(restoreCPSR);

        if (switchThumb)
            MOV(W1, addr);
@ -316,11 +297,11 @@ void Compiler::Comp_JumpTo(Arm64Gen::ARM64Reg addr, bool switchThumb, bool resto
        MOV(X0, RCPU);
        MOVI2R(W2, restoreCPSR);
        if (Num == 0)
-            QuickCallFunction(X3, JumpToTrampoline<ARMv5>);
+            QuickCallFunction(X3, jumpToTrampoline<ARMv5>);
        else
-            QuickCallFunction(X3, JumpToTrampoline<ARMv4>);
+            QuickCallFunction(X3, jumpToTrampoline<ARMv4>);

-        PopRegs(restoreCPSR, true);
+        PopRegs(restoreCPSR);
        LoadCycles();
        LoadCPSR();
        if (CurInstr.Cond() < 0xE)
@ -389,7 +370,7 @@ void Compiler::T_Comp_BranchXchangeReg()
    {
        if (Num == 1)
        {
-            Log(LogLevel::Warn, "BLX unsupported on ARM7!!!\n");
+            printf("BLX unsupported on ARM7!!!\n");
            return;
        }
        MOV(W0, MapReg(CurInstr.A_Reg(3)));
@ -437,4 +418,4 @@ void Compiler::T_Comp_BL_Merged()
    Comp_JumpTo(target);
 }

-}
+}
--- a/src/ARMJIT_A64/ARMJIT_Compiler.cpp
+++ b/src/ARMJIT_A64/ARMJIT_Compiler.cpp
@ -1,28 +1,17 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARMJIT_Compiler.h"

 #include "../ARMJIT_Internal.h"
 #include "../ARMInterpreter.h"
-#include "../ARMJIT.h"
-#include "../NDS.h"
-#include "../ARMJIT_Global.h"
+#include "../Config.h"
+
+#ifdef __SWITCH__
+#include <switch.h>
+
+extern char __start__;
+#else
+#include <sys/mman.h>
+#include <unistd.h>
+#endif

 #include <stdlib.h>

@ -30,7 +19,7 @@ using namespace Arm64Gen;

 extern "C" void ARM_Ret();

-namespace melonDS
+namespace ARMJIT
 {

 /*
@ -47,14 +36,14 @@ namespace melonDS

 template <>
 const ARM64Reg RegisterCache<Compiler, ARM64Reg>::NativeRegAllocOrder[] =
-{
-    W19, W20, W21, W22, W23, W24, W25,
-    W8, W9, W10, W11, W12, W13, W14, W15
-};
+    {W19, W20, W21, W22, W23, W24, W25, W26};
 template <>
-const int RegisterCache<Compiler, ARM64Reg>::NativeRegsAvailable = 15;
+const int RegisterCache<Compiler, ARM64Reg>::NativeRegsAvailable = 8;

-const BitSet32 CallerSavedPushRegs({W8, W9, W10, W11, W12, W13, W14, W15});
+const int JitMemSize = 16 * 1024 * 1024;
+#ifndef __SWITCH__
+u8 JitMem[JitMemSize];
+#endif

 void Compiler::MovePC()
 {
@ -92,7 +81,7 @@ void Compiler::A_Comp_MSR()
    if (CurInstr.Instr & (1 << 25))
    {
        val = W0;
-        MOVI2R(val, melonDS::ROR((CurInstr.Instr & 0xFF), ((CurInstr.Instr >> 7) & 0x1E)));
+        MOVI2R(val, ::ROR((CurInstr.Instr & 0xFF), ((CurInstr.Instr >> 7) & 0x1E)));
    }
    else
    {
@ -153,67 +142,52 @@ void Compiler::A_Comp_MSR()
            MOV(W2, RCPSR);
            MOV(X0, RCPU);

-            PushRegs(true, true);
-            QuickCallFunction(X3, UpdateModeTrampoline);
-            PopRegs(true, true);
+            PushRegs(true);
+
+            QuickCallFunction(X3, (void*)&UpdateModeTrampoline);
+        
+            PopRegs(true);
        }
    }
 }

-
-void Compiler::PushRegs(bool saveHiRegs, bool saveRegsToBeChanged, bool allowUnload)
+void Compiler::PushRegs(bool saveHiRegs)
 {
-    BitSet32 loadedRegs(RegCache.LoadedRegs);
-
    if (saveHiRegs)
    {
-        BitSet32 hiRegsLoaded(RegCache.LoadedRegs & 0x7F00);
-        for (int reg : hiRegsLoaded)
+        if (Thumb || CurInstr.Cond() == 0xE)
        {
-            if (Thumb || CurInstr.Cond() == 0xE)
+            BitSet16 hiRegsLoaded(RegCache.LoadedRegs & 0x7F00);
+            for (int reg : hiRegsLoaded)
                RegCache.UnloadRegister(reg);
-            else
-                SaveReg(reg, RegCache.Mapping[reg]);
-            // prevent saving the register twice
-            loadedRegs[reg] = false;
        }
-    }
-
-    for (int reg : loadedRegs)
-    {
-        if (CallerSavedPushRegs[RegCache.Mapping[reg]]
-            && (saveRegsToBeChanged || !((1<<reg) & CurInstr.Info.DstRegs && !((1<<reg) & CurInstr.Info.SrcRegs))))
+        else
        {
-            if ((Thumb || CurInstr.Cond() == 0xE) && !((1 << reg) & (CurInstr.Info.DstRegs|CurInstr.Info.SrcRegs)) && allowUnload)
-                RegCache.UnloadRegister(reg);
-            else
+            BitSet16 hiRegsDirty(RegCache.LoadedRegs & 0x7F00);
+            for (int reg : hiRegsDirty)
                SaveReg(reg, RegCache.Mapping[reg]);
        }
    }
 }

-void Compiler::PopRegs(bool saveHiRegs, bool saveRegsToBeChanged)
+void Compiler::PopRegs(bool saveHiRegs)
 {
-    BitSet32 loadedRegs(RegCache.LoadedRegs);
-    for (int reg : loadedRegs)
+    if (saveHiRegs)
    {
-        if ((saveHiRegs && reg >= 8 && reg < 15)
-            || (CallerSavedPushRegs[RegCache.Mapping[reg]]
-                && (saveRegsToBeChanged || !((1<<reg) & CurInstr.Info.DstRegs && !((1<<reg) & CurInstr.Info.SrcRegs)))))
-        {
+        BitSet16 hiRegsLoaded(RegCache.LoadedRegs & 0x7F00);
+
+        for (int reg : hiRegsLoaded)
            LoadReg(reg, RegCache.Mapping[reg]);
-        }
    }
 }

-Compiler::Compiler(melonDS::NDS& nds) : Arm64Gen::ARM64XEmitter(), NDS(nds)
+Compiler::Compiler()
 {
 #ifdef __SWITCH__
    JitRWBase = aligned_alloc(0x1000, JitMemSize);

    JitRXStart = (u8*)&__start__ - JitMemSize - 0x1000;
-    virtmemLock();
-    JitRWStart = virtmemFindAslr(JitMemSize, 0x1000);
+    JitRWStart = virtmemReserve(JitMemSize);
    MemoryInfo info = {0};
    u32 pageInfo = {0};
    int i = 0;
@ -226,7 +200,7 @@ Compiler::Compiler(melonDS::NDS& nds) : Arm64Gen::ARM64XEmitter(), NDS(nds)
            break;
        if (i++ > 8)
        {
-            Log(LogLevel::Error, "couldn't find unmapped place for jit memory\n");
+            printf("couldn't find unmapped place for jit memory\n");
            JitRXStart = NULL;
        }
    }
@ -240,18 +214,16 @@ Compiler::Compiler(melonDS::NDS& nds) : Arm64Gen::ARM64XEmitter(), NDS(nds)
    succeded = R_SUCCEEDED(svcMapProcessMemory(JitRWStart, envGetOwnProcessHandle(), (u64)JitRXStart, JitMemSize));
    assert(succeded);

-    virtmemUnlock();
-
    SetCodeBase((u8*)JitRWStart, (u8*)JitRXStart);
    JitMemMainSize = JitMemSize;
 #else
-    ARMJIT_Global::Init();
+    u64 pageSize = sysconf(_SC_PAGE_SIZE);
+    u8* pageAligned = (u8*)(((u64)JitMem & ~(pageSize - 1)) + pageSize);
+    u64 alignedSize = (((u64)JitMem + sizeof(JitMem)) & ~(pageSize - 1)) - (u64)pageAligned;
+    mprotect(pageAligned, alignedSize, PROT_EXEC | PROT_READ | PROT_WRITE);

-    CodeMemBase = ARMJIT_Global::AllocateCodeMem();
-    nds.JIT.JitEnableWrite();
-
-    SetCodeBase(reinterpret_cast<u8*>(CodeMemBase), reinterpret_cast<u8*>(CodeMemBase));
-    JitMemMainSize = ARMJIT_Global::CodeMemorySliceSize;
+    SetCodeBase(pageAligned, pageAligned);
+    JitMemMainSize = alignedSize;
 #endif
    SetCodePtr(0);

@ -262,7 +234,6 @@ Compiler::Compiler(melonDS::NDS& nds) : Arm64Gen::ARM64XEmitter(), NDS(nds)
    }

    /*
-        W4 - whether the register was written to
        W5 - mode
        W1 - reg num
        W3 - in/out value of reg
@ -352,11 +323,9 @@ Compiler::Compiler(melonDS::NDS& nds) : Arm64Gen::ARM64XEmitter(), NDS(nds)
        {
            for (int size = 0; size < 3; size++)
            {
-                for (int reg = 0; reg < 32; reg++)
+                for (int reg = 0; reg < 8; reg++)
                {
-                    if (!(reg == W4 || (reg >= W8 && reg <= W15) || (reg >= W19 && reg <= W25)))
-                        continue;
-                    ARM64Reg rdMapped = (ARM64Reg)reg;
+                    ARM64Reg rdMapped = (ARM64Reg)(W19 + reg);
                    PatchedStoreFuncs[consoleType][num][size][reg] = GetRXPtr();
                    if (num == 0)
                    {
@ -367,7 +336,7 @@ Compiler::Compiler(melonDS::NDS& nds) : Arm64Gen::ARM64XEmitter(), NDS(nds)
                    {
                        MOV(W1, rdMapped);
                    }
-                    ABI_PushRegisters(BitSet32({30}) | CallerSavedPushRegs);
+                    ABI_PushRegisters({30});
                    if (consoleType == 0)
                    {
                        switch ((8 << size) |  num)
@ -393,7 +362,7 @@ Compiler::Compiler(melonDS::NDS& nds) : Arm64Gen::ARM64XEmitter(), NDS(nds)
                        }
                    }
                    
-                    ABI_PopRegisters(BitSet32({30}) | CallerSavedPushRegs);
+                    ABI_PopRegisters({30});
                    RET();

                    for (int signextend = 0; signextend < 2; signextend++)
@ -401,7 +370,7 @@ Compiler::Compiler(melonDS::NDS& nds) : Arm64Gen::ARM64XEmitter(), NDS(nds)
                        PatchedLoadFuncs[consoleType][num][size][signextend][reg] = GetRXPtr();
                        if (num == 0)
                            MOV(X1, RCPU);
-                        ABI_PushRegisters(BitSet32({30}) | CallerSavedPushRegs);
+                        ABI_PushRegisters({30});
                        if (consoleType == 0)
                        {
                            switch ((8 << size) |  num)
@ -426,7 +395,7 @@ Compiler::Compiler(melonDS::NDS& nds) : Arm64Gen::ARM64XEmitter(), NDS(nds)
                            case 9: QuickCallFunction(X3, SlowRead7<u8, 1>); break;
                            }
                        }
-                        ABI_PopRegisters(BitSet32({30}) | CallerSavedPushRegs);
+                        ABI_PopRegisters({30});
                        if (size == 32)
                            MOV(rdMapped, W0);
                        else if (signextend)
@ -457,14 +426,12 @@ Compiler::~Compiler()
    {
        bool succeded = R_SUCCEEDED(svcUnmapProcessMemory(JitRWStart, envGetOwnProcessHandle(), (u64)JitRXStart, JitMemSize));
        assert(succeded);
+        virtmemFree(JitRWStart, JitMemSize);
        succeded = R_SUCCEEDED(svcUnmapProcessCodeMemory(envGetOwnProcessHandle(), (u64)JitRXStart, (u64)JitRWBase, JitMemSize));
        assert(succeded);
        free(JitRWBase);
    }
 #endif
-
-    ARMJIT_Global::FreeCodeMem(CodeMemBase);
-    ARMJIT_Global::DeInit();
 }

 void Compiler::LoadCycles()
@ -482,12 +449,12 @@ void Compiler::LoadReg(int reg, ARM64Reg nativeReg)
    if (reg == 15)
        MOVI2R(nativeReg, R15);
    else
-        LDR(INDEX_UNSIGNED, nativeReg, RCPU, offsetof(ARM, R) + reg*4);
+        LDR(INDEX_UNSIGNED, nativeReg, RCPU, offsetof(ARM, R[reg]));
 }

 void Compiler::SaveReg(int reg, ARM64Reg nativeReg)
 {
-    STR(INDEX_UNSIGNED, nativeReg, RCPU, offsetof(ARM, R) + reg*4);
+    STR(INDEX_UNSIGNED, nativeReg, RCPU, offsetof(ARM, R[reg]));
 }

 void Compiler::LoadCPSR()
@ -667,23 +634,22 @@ void Compiler::Comp_BranchSpecialBehaviour(bool taken)
    {
        RegCache.PrepareExit();

-        if (ConstantCycles)
-            ADD(RCycles, RCycles, ConstantCycles);
+        ADD(RCycles, RCycles, ConstantCycles);
        QuickTailCall(X0, ARM_Ret);
    }
 }

-JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[], int instrsCount, bool hasMemInstr)
+JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[], int instrsCount)
 {
    if (JitMemMainSize - GetCodeOffset() < 1024 * 16)
    {
-        Log(LogLevel::Debug, "JIT near memory full, resetting...\n");
-        NDS.JIT.ResetBlockCache();
+        printf("JIT near memory full, resetting...\n");
+        ResetBlockCache();
    }
    if ((JitMemMainSize +  JitMemSecondarySize) - OtherCodeRegion < 1024 * 8)
    {
-        Log(LogLevel::Debug, "JIT far memory full, resetting...\n");
-        NDS.JIT.ResetBlockCache();
+        printf("JIT far memory full, resetting...\n");
+        ResetBlockCache();
    }

    JitBlockEntry res = (JitBlockEntry)GetRXPtr();
@ -695,9 +661,6 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]
    RegCache = RegisterCache<Compiler, ARM64Reg>(this, instrs, instrsCount, true);
    CPSRDirty = false;

-    if (hasMemInstr)
-        MOVP2R(RMemBase, Num == 0 ? NDS.JIT.Memory.FastMem9Start : NDS.JIT.Memory.FastMem7Start);
-
    for (int i = 0; i < instrsCount; i++)
    {
        CurInstr = instrs[i];
@ -746,9 +709,7 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]
                QuickCallFunction(X1, InterpretTHUMB[CurInstr.Info.Kind]);
            }
            else
-            {
                (this->*comp)();
-            }
        }
        else
        {
@ -764,12 +725,10 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]
                }
            }
            else if (cond == 0xF)
-            {
                Comp_AddCycles_C();
-            }
            else
            {
-                IrregularCycles = comp == NULL;
+                IrregularCycles = false;

                FixupBranch skipExecute;
                if (cond < 0xE)
@ -794,17 +753,14 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]
                        FixupBranch skipNop = B();
                        SetJumpTarget(skipExecute);

-                        if (IrregularCycles)
-                            Comp_AddCycles_C(true);
+                        Comp_AddCycles_C();

                        Comp_BranchSpecialBehaviour(false);

                        SetJumpTarget(skipNop);
                    }
                    else
-                    {
                        SetJumpTarget(skipExecute);
-                    }
                }

            }
@ -819,8 +775,7 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]

    RegCache.Flush();

-    if (ConstantCycles)
-        ADD(RCycles, RCycles, ConstantCycles);
+    ADD(RCycles, RCycles, ConstantCycles);
    QuickTailCall(X0, ARM_Ret);

    FlushIcache();
@ -844,7 +799,7 @@ void Compiler::Reset()
 void Compiler::Comp_AddCycles_C(bool forceNonConstant)
 {
    s32 cycles = Num ?
-        NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 1 : 3]
+        NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 1 : 3]
        : ((R15 & 0x2) ? 0 : CurInstr.CodeCycles);

    if (forceNonConstant)
@ -858,7 +813,7 @@ void Compiler::Comp_AddCycles_CI(u32 numI)
    IrregularCycles = true;

    s32 cycles = (Num ?
-        NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2]
+        NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2]
        : ((R15 & 0x2) ? 0 : CurInstr.CodeCycles)) + numI;

    if (Thumb || CurInstr.Cond() == 0xE)
@ -872,7 +827,7 @@ void Compiler::Comp_AddCycles_CI(u32 c, ARM64Reg numI, ArithOption shift)
    IrregularCycles = true;

    s32 cycles = (Num ?
-        NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2]
+        NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2]
        : ((R15 & 0x2) ? 0 : CurInstr.CodeCycles)) + c;

    ADD(RCycles, RCycles, cycles);
@ -892,7 +847,7 @@ void Compiler::Comp_AddCycles_CDI()

        s32 cycles;

-        s32 numC = NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2];
+        s32 numC = NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2];
        s32 numD = CurInstr.DataCycles;

        if ((CurInstr.DataRegion >> 24) == 0x02) // mainRAM
@ -937,7 +892,7 @@ void Compiler::Comp_AddCycles_CD()
    }
    else
    {
-        s32 numC = NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2];
+        s32 numC = NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2];
        s32 numD = CurInstr.DataCycles;

        if ((CurInstr.DataRegion >> 24) == 0x02)
--- a/src/ARMJIT_A64/ARMJIT_Compiler.h
+++ b/src/ARMJIT_A64/ARMJIT_Compiler.h
@ -1,27 +1,8 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #ifndef ARMJIT_A64_COMPILER_H
 #define ARMJIT_A64_COMPILER_H

-#if defined(JIT_ENABLED) && defined(__aarch64__)
-
 #include "../ARM.h"
+#include "../ARMJIT.h"

 #include "../dolphin/Arm64Emitter.h"

@ -30,10 +11,9 @@

 #include <unordered_map>

-namespace melonDS
+namespace ARMJIT
 {
-class ARMJIT;
-const Arm64Gen::ARM64Reg RMemBase = Arm64Gen::X26;
+
 const Arm64Gen::ARM64Reg RCPSR = Arm64Gen::W27;
 const Arm64Gen::ARM64Reg RCycles = Arm64Gen::W28;
 const Arm64Gen::ARM64Reg RCPU = Arm64Gen::X29;
@ -69,7 +49,7 @@ struct Op2
    bool IsSimpleReg()
    { return !IsImm && !Reg.ShiftAmount && Reg.ShiftType == Arm64Gen::ST_LSL; }
    bool ImmFits12Bit()
-    { return IsImm && ((Imm & 0xFFF) == Imm); }
+    { return IsImm && (Imm & 0xFFF == Imm); }
    bool IsZero()
    { return IsImm && !Imm; }

@ -98,11 +78,11 @@ class Compiler : public Arm64Gen::ARM64XEmitter
 public:
    typedef void (Compiler::*CompileFunc)();

-    explicit Compiler(melonDS::NDS& nds);
-    ~Compiler() override;
+    Compiler();
+    ~Compiler();

-    void PushRegs(bool saveHiRegs, bool saveRegsToBeChanged, bool allowUnload = true);
-    void PopRegs(bool saveHiRegs, bool saveRegsToBeChanged);
+    void PushRegs(bool saveHiRegs);
+    void PopRegs(bool saveHiRegs);

    Arm64Gen::ARM64Reg MapReg(int reg)
    {
@ -110,11 +90,11 @@ public:
        return RegCache.Mapping[reg];
    }

-    JitBlockEntry CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[], int instrsCount, bool hasMemInstr);
+    JitBlockEntry CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[], int instrsCount);

    bool CanCompile(bool thumb, u16 kind);

-    bool FlagsNZNeeded() const
+    bool FlagsNZNeeded()
    {
        return CurInstr.SetFlags & 0xC;
    }
@ -186,7 +166,7 @@ public:
    void T_Comp_BL_LONG_2();
    void T_Comp_BL_Merged();

-    s32 Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc, bool decrement, bool usermode, bool skipLoadingRn);
+    s32 Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc, bool decrement, bool usermode);

    void Comp_Mul_Mla(bool S, bool mla, Arm64Gen::ARM64Reg rd, Arm64Gen::ARM64Reg rm, Arm64Gen::ARM64Reg rs, Arm64Gen::ARM64Reg rn);

@ -234,7 +214,7 @@ public:
        return (u8*)entry - GetRXBase();
    }

-    bool IsJITFault(const u8* pc);
+    bool IsJITFault(u8* pc);
    u8* RewriteMemAccess(u8* pc);

    void SwapCodeRegion()
@ -244,7 +224,6 @@ public:
        OtherCodeRegion = offset;
    }

-    melonDS::NDS& NDS;
    ptrdiff_t OtherCodeRegion;

    bool Exit;
@ -262,8 +241,17 @@ public:
    u32 JitMemSecondarySize;
    u32 JitMemMainSize;

+    void* ReadBanked, *WriteBanked;
+
+    void* JumpToFuncs9[3];
+    void* JumpToFuncs7[3];
+
    std::unordered_map<ptrdiff_t, LoadStorePatch> LoadStorePatches; 

+    // [Console Type][Num][Size][Sign Extend][Output register]
+    void* PatchedLoadFuncs[2][2][3][2][8];
+    void* PatchedStoreFuncs[2][2][3][8];
+
    RegisterCache<Compiler, Arm64Gen::ARM64Reg> RegCache;

    bool CPSRDirty = false;
@ -275,20 +263,8 @@ public:
    void* JitRWStart;
    void* JitRXStart;
 #endif
-    void* CodeMemBase;
-
-    void* ReadBanked, *WriteBanked;
-
-    void* JumpToFuncs9[3];
-    void* JumpToFuncs7[3];
-
-    // [Console Type][Num][Size][Sign Extend][Output register]
-    void* PatchedLoadFuncs[2][2][3][2][32];
-    void* PatchedStoreFuncs[2][2][3][32];
 };

 }

 #endif
-
-#endif
--- a/src/ARMJIT_A64/ARMJIT_Linkage.S
+++ b/src/ARMJIT_A64/ARMJIT_Linkage.S
@ -1,21 +1,3 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "../ARMJIT_x64/ARMJIT_Offsets.h"

 .text
@ -94,8 +76,3 @@ ARM_RestoreContext:
    mov sp, x17

    br x18
-
-#if !defined(__APPLE__) && !defined(__WIN32__)
-.section .note.GNU-stack,"",@progbits
-#endif
-
--- a/src/ARMJIT_A64/ARMJIT_LoadStore.cpp
+++ b/src/ARMJIT_A64/ARMJIT_LoadStore.cpp
@ -1,34 +1,15 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARMJIT_Compiler.h"

-#include "../ARMJIT.h"
+#include "../Config.h"

 #include "../ARMJIT_Memory.h"
-#include "../NDS.h"

 using namespace Arm64Gen;

-namespace melonDS
+namespace ARMJIT
 {

-bool Compiler::IsJITFault(const u8* pc)
+bool Compiler::IsJITFault(u8* pc)
 {
    return (u64)pc >= (u64)GetRXBase() && (u64)pc - (u64)GetRXBase() < (JitMemMainSize + JitMemSecondarySize);
 }
@ -57,17 +38,18 @@ u8* Compiler::RewriteMemAccess(u8* pc)

        return pc + (ptrdiff_t)patch.PatchOffset;
    }
-    Log(LogLevel::Error, "this is a JIT bug! %08x\n", __builtin_bswap32(*(u32*)pc));
+    printf("this is a JIT bug! %08x\n", __builtin_bswap32(*(u32*)pc));
    abort();
 }

 bool Compiler::Comp_MemLoadLiteral(int size, bool signExtend, int rd, u32 addr)
 {
-    u32 localAddr = NDS.JIT.LocaliseCodeAddress(Num, addr);
+    u32 localAddr = LocaliseCodeAddress(Num, addr);

-    int invalidLiteralIdx = NDS.JIT.InvalidLiterals.Find(localAddr);
+    int invalidLiteralIdx = InvalidLiterals.Find(localAddr);
    if (invalidLiteralIdx != -1)
    {
+        InvalidLiterals.Remove(invalidLiteralIdx);
        return false;
    }

@ -80,7 +62,7 @@ bool Compiler::Comp_MemLoadLiteral(int size, bool signExtend, int rd, u32 addr)
    if (size == 32)
    {
        CurCPU->DataRead32(addr & ~0x3, &val);
-        val = melonDS::ROR(val, (addr & 0x3) << 3);
+        val = ::ROR(val, (addr & 0x3) << 3);
    }
    else if (size == 16)
    {
@ -112,7 +94,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
    if (size == 16)
        addressMask = ~1;

-    if (NDS.JIT.LiteralOptimizationsEnabled() && rn == 15 && rd != 15 && offset.IsImm && !(flags & (memop_Post|memop_Store|memop_Writeback)))
+    if (Config::JIT_LiteralOptimisations && rn == 15 && rd != 15 && offset.IsImm && !(flags & (memop_Post|memop_Store|memop_Writeback)))
    {
        u32 addr = R15 + offset.Imm * ((flags & memop_SubtractOffset) ? -1 : 1);
        
@ -134,12 +116,6 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
        rnMapped = W3;
    }

-    if (flags & memop_Store && flags & (memop_Post|memop_Writeback) && rd == rn)
-    {
-        MOV(W4, rdMapped);
-        rdMapped = W4;
-    }
-
    ARM64Reg finalAddr = W0;
    if (flags & memop_Post)
    {
@ -147,7 +123,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
        MOV(W0, rnMapped);
    }

-    bool addrIsStatic = NDS.JIT.LiteralOptimizationsEnabled()
+    bool addrIsStatic = Config::JIT_LiteralOptimisations
        && RegCache.IsLiteral(rn) && offset.IsImm && !(flags & (memop_Writeback|memop_Post));
    u32 staticAddress;
    if (addrIsStatic)
@ -186,18 +162,20 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
        MOV(rnMapped, W0);

    u32 expectedTarget = Num == 0
-        ? NDS.JIT.Memory.ClassifyAddress9(addrIsStatic ? staticAddress : CurInstr.DataRegion)
-        : NDS.JIT.Memory.ClassifyAddress7(addrIsStatic ? staticAddress : CurInstr.DataRegion);
+        ? ARMJIT_Memory::ClassifyAddress9(addrIsStatic ? staticAddress : CurInstr.DataRegion)
+        : ARMJIT_Memory::ClassifyAddress7(addrIsStatic ? staticAddress : CurInstr.DataRegion);

-    if (NDS.JIT.FastMemoryEnabled() && ((!Thumb && CurInstr.Cond() != 0xE) || NDS.JIT.Memory.IsFastmemCompatible(expectedTarget)))
+    if (Config::JIT_FastMemory && ((!Thumb && CurInstr.Cond() != 0xE) || ARMJIT_Memory::IsFastmemCompatible(expectedTarget)))
    {
        ptrdiff_t memopStart = GetCodeOffset();
        LoadStorePatch patch;

-        assert((rdMapped >= W8 && rdMapped <= W15) || (rdMapped >= W19 && rdMapped <= W25) || rdMapped == W4);
        patch.PatchFunc = flags & memop_Store
-            ? PatchedStoreFuncs[NDS.ConsoleType][Num][__builtin_ctz(size) - 3][rdMapped]
-            : PatchedLoadFuncs[NDS.ConsoleType][Num][__builtin_ctz(size) - 3][!!(flags & memop_SignExtend)][rdMapped];
+            ? PatchedStoreFuncs[NDS::ConsoleType][Num][__builtin_ctz(size) - 3][rdMapped - W19]
+            : PatchedLoadFuncs[NDS::ConsoleType][Num][__builtin_ctz(size) - 3][!!(flags & memop_SignExtend)][rdMapped - W19];
+        assert(rdMapped - W19 >= 0 && rdMapped - W19 < 8);
+
+        MOVP2R(X7, Num == 0 ? ARMJIT_Memory::FastMem9Start : ARMJIT_Memory::FastMem7Start);

        // take a chance at fastmem
        if (size > 8)
@ -206,11 +184,11 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
        ptrdiff_t loadStorePosition = GetCodeOffset();
        if (flags & memop_Store)
        {
-            STRGeneric(size, rdMapped, size > 8 ? X1 : X0, RMemBase);
+            STRGeneric(size, rdMapped, size > 8 ? X1 : X0, X7);
        }
        else
        {
-            LDRGeneric(size, flags & memop_SignExtend, rdMapped, size > 8 ? X1 : X0, RMemBase);
+            LDRGeneric(size, flags & memop_SignExtend, rdMapped, size > 8 ? X1 : X0, X7);
            if (size == 32 && !addrIsStatic)
            {
                UBFIZ(W0, W0, 3, 2);
@ -226,9 +204,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
    {
        void* func = NULL;
        if (addrIsStatic)
-            func = NDS.JIT.Memory.GetFuncForAddr(CurCPU, staticAddress, flags & memop_Store, size);
-
-        PushRegs(false, false);
+            func = ARMJIT_Memory::GetFuncForAddr(CurCPU, staticAddress, flags & memop_Store, size);

        if (func)
        {
@ -236,8 +212,6 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
                MOV(W1, rdMapped);
            QuickCallFunction(X2, (void (*)())func);

-            PopRegs(false, false);
-
            if (!(flags & memop_Store))
            {
                if (size == 32)
@ -264,7 +238,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
                if (flags & memop_Store)
                {
                    MOV(W2, rdMapped);
-                    switch (size | NDS.ConsoleType)
+                    switch (size | NDS::ConsoleType)
                    {
                    case 32: QuickCallFunction(X3, SlowWrite9<u32, 0>); break;
                    case 33: QuickCallFunction(X3, SlowWrite9<u32, 1>); break;
@ -276,7 +250,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
                }
                else
                {
-                    switch (size | NDS.ConsoleType)
+                    switch (size | NDS::ConsoleType)
                    {
                    case 32: QuickCallFunction(X3, SlowRead9<u32, 0>); break;
                    case 33: QuickCallFunction(X3, SlowRead9<u32, 1>); break;
@ -292,7 +266,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
                if (flags & memop_Store)
                {
                    MOV(W1, rdMapped);
-                    switch (size | NDS.ConsoleType)
+                    switch (size | NDS::ConsoleType)
                    {
                    case 32: QuickCallFunction(X3, SlowWrite7<u32, 0>); break;
                    case 33: QuickCallFunction(X3, SlowWrite7<u32, 1>); break;
@ -304,7 +278,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
                }
                else
                {
-                    switch (size | NDS.ConsoleType)
+                    switch (size | NDS::ConsoleType)
                    {
                    case 32: QuickCallFunction(X3, SlowRead7<u32, 0>); break;
                    case 33: QuickCallFunction(X3, SlowRead7<u32, 1>); break;
@ -316,8 +290,6 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
                }
            }

-            PopRegs(false, false);
-
            if (!(flags & memop_Store))
            {
                if (size == 32)
@ -333,7 +305,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, Op2 offset, int size, int flags)
    if (CurInstr.Info.Branches())
    {
        if (size < 32)
-            Log(LogLevel::Debug, "LDR size < 32 branching?\n");
+            printf("LDR size < 32 branching?\n");
        Comp_JumpTo(rdMapped, Num == 0, false);
    }
 }
@ -453,7 +425,7 @@ void Compiler::T_Comp_LoadPCRel()
    u32 offset = ((CurInstr.Instr & 0xFF) << 2);
    u32 addr = (R15 & ~0x2) + offset;

-    if (!NDS.JIT.LiteralOptimizationsEnabled() || !Comp_MemLoadLiteral(32, false, CurInstr.T_Reg(8), addr))
+    if (!Config::JIT_LiteralOptimisations || !Comp_MemLoadLiteral(32, false, CurInstr.T_Reg(8), addr))
        Comp_MemAccess(CurInstr.T_Reg(8), 15, Op2(offset), 32, 0);
 }

@ -465,7 +437,7 @@ void Compiler::T_Comp_MemSPRel()
    Comp_MemAccess(CurInstr.T_Reg(8), 13, Op2(offset), 32, load ? 0 : memop_Store);
 }

-s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc, bool decrement, bool usermode, bool skipLoadingRn)
+s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc, bool decrement, bool usermode)
 {
    IrregularCycles = true;

@ -474,8 +446,7 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
    if (regsCount == 0)
        return 0; // actually not the right behaviour TODO: fix me

-    int firstReg = *regs.begin();
-    if (regsCount == 1 && !usermode && RegCache.LoadedRegs & (1 << firstReg) && !(firstReg == rn && skipLoadingRn))
+    if (regsCount == 1 && !usermode && RegCache.LoadedRegs & (1 << *regs.begin()))
    {
        int flags = 0;
        if (store)
@ -484,7 +455,7 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
            flags |= memop_SubtractOffset;
        Op2 offset = preinc ? Op2(4) : Op2(0);

-        Comp_MemAccess(firstReg, rn, offset, 32, flags);
+        Comp_MemAccess(*regs.begin(), rn, offset, 32, flags);

        return decrement ? -4 : 4;
    }
@ -495,32 +466,40 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
        Comp_AddCycles_CDI();

    int expectedTarget = Num == 0
-        ? NDS.JIT.Memory.ClassifyAddress9(CurInstr.DataRegion)
-        : NDS.JIT.Memory.ClassifyAddress7(CurInstr.DataRegion);
+        ? ARMJIT_Memory::ClassifyAddress9(CurInstr.DataRegion)
+        : ARMJIT_Memory::ClassifyAddress7(CurInstr.DataRegion);

-    bool compileFastPath = NDS.JIT.FastMemoryEnabled()
-        && store && !usermode && (CurInstr.Cond() < 0xE || NDS.JIT.Memory.IsFastmemCompatible(expectedTarget));
+    bool compileFastPath = Config::JIT_FastMemory
+        && store && !usermode && (CurInstr.Cond() < 0xE || ARMJIT_Memory::IsFastmemCompatible(expectedTarget));

+    if (decrement)
    {
-        s32 offset = decrement
-            ? -regsCount * 4 + (preinc ? 0 : 4)
-            : (preinc ? 4 : 0);
-
+        s32 offset = -regsCount * 4 + (preinc ? 0 : 4);
        if (offset)
+        {
            ADDI2R(W0, MapReg(rn), offset);
-        else if (compileFastPath)
-            ANDI2R(W0, MapReg(rn), ~3);
+            ANDI2R(W0, W0, ~3);
+        }
        else
-            MOV(W0, MapReg(rn));
+        {
+            ANDI2R(W0, MapReg(rn), ~3);
+        }
+    }
+    else
+    {
+        ANDI2R(W0, MapReg(rn), ~3);
+        if (preinc)
+            ADD(W0, W0, 4);
    }

    u8* patchFunc;
    if (compileFastPath)
    {
        ptrdiff_t fastPathStart = GetCodeOffset();
-        ptrdiff_t loadStoreOffsets[8];
+        ptrdiff_t loadStoreOffsets[16];

-        ADD(X1, RMemBase, X0);
+        MOVP2R(X1, Num == 0 ? ARMJIT_Memory::FastMem9Start : ARMJIT_Memory::FastMem7Start);
+        ADD(X1, X1, X0);

        u32 offset = 0;
        BitSet16::Iterator it = regs.begin();
@ -540,16 +519,12 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
            loadStoreOffsets[i++] = GetCodeOffset();

            if (store)
-            {
                STR(INDEX_UNSIGNED, first, X1, offset);
-            }
-            else if (!(reg == rn && skipLoadingRn))
-            {
+            else
                LDR(INDEX_UNSIGNED, first, X1, offset);

-                if (!(RegCache.LoadedRegs & (1 << reg)))
-                    SaveReg(reg, first);
-            }
+            if (!(RegCache.LoadedRegs & (1 << reg)) && !store)
+                SaveReg(reg, first);

            offset += 4;
        }
@ -563,38 +538,25 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc

            ARM64Reg first = W3, second = W4;
            if (RegCache.LoadedRegs & (1 << reg))
-            {
-                if (!(reg == rn && skipLoadingRn))
-                    first = MapReg(reg);
-            }
+                first = MapReg(reg);
            else if (store)
-            {
                LoadReg(reg, first);
-            }
            if (RegCache.LoadedRegs & (1 << nextReg))
-            {
-                if (!(nextReg == rn && skipLoadingRn))
-                    second = MapReg(nextReg);
-            }
+                second = MapReg(nextReg);
            else if (store)
-            {
                LoadReg(nextReg, second);
-            }

            loadStoreOffsets[i++] = GetCodeOffset();
+
            if (store)
-            {
                STP(INDEX_SIGNED, first, second, X1, offset);
-            }
            else
-            {
                LDP(INDEX_SIGNED, first, second, X1, offset);
-            
-                if (!(RegCache.LoadedRegs & (1 << reg)))
-                    SaveReg(reg, first);
-                if (!(RegCache.LoadedRegs & (1 << nextReg)))
-                    SaveReg(nextReg, second);
-            }
+
+            if (!(RegCache.LoadedRegs & (1 << reg)) && !store)
+                SaveReg(reg, first);
+            if (!(RegCache.LoadedRegs & (1 << nextReg)) && !store)
+                SaveReg(nextReg, second);

            offset += 8;
        }
@ -604,8 +566,7 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
        SwapCodeRegion();
        patchFunc = (u8*)GetRXPtr();
        patch.PatchFunc = patchFunc;
-        u32 numLoadStores = i;
-        for (i = 0; i < numLoadStores; i++)
+        for (i = 0; i < regsCount; i++)
        {
            patch.PatchOffset = fastPathStart - loadStoreOffsets[i];
            LoadStorePatches[loadStoreOffsets[i]] = patch;
@ -673,15 +634,13 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
        }
    }

-    PushRegs(false, false, !compileFastPath);
-
    ADD(X1, SP, 0);
    MOVI2R(W2, regsCount);

    if (Num == 0)
    {
        MOV(X3, RCPU);
-        switch ((u32)store * 2 | NDS.ConsoleType)
+        switch ((u32)store * 2 | NDS::ConsoleType)
        {
        case 0: QuickCallFunction(X4, SlowBlockTransfer9<false, 0>); break;
        case 1: QuickCallFunction(X4, SlowBlockTransfer9<false, 1>); break;
@ -691,7 +650,7 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
    }
    else
    {
-        switch ((u32)store * 2 | NDS.ConsoleType)
+        switch ((u32)store * 2 | NDS::ConsoleType)
        {
        case 0: QuickCallFunction(X4, SlowBlockTransfer7<false, 0>); break;
        case 1: QuickCallFunction(X4, SlowBlockTransfer7<false, 1>); break;
@ -700,8 +659,6 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
        }
    }

-    PopRegs(false, false);
-
    if (!store)
    {
        if (usermode && !regs[15] && (regs & BitSet16(0x7f00)))
@ -720,23 +677,20 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
                LDR(INDEX_UNSIGNED, W3, SP, i * 8);
                MOVI2R(W1, reg - 8);
                BL(WriteBanked);
-                if (!(reg == rn && skipLoadingRn))
-                {
-                    FixupBranch alreadyWritten = CBNZ(W4);
-                    if (RegCache.LoadedRegs & (1 << reg))
-                        MOV(MapReg(reg), W3);
-                    else
-                        SaveReg(reg, W3);
-                    SetJumpTarget(alreadyWritten);
-                }
+                FixupBranch alreadyWritten = CBNZ(W4);
+                if (RegCache.LoadedRegs & (1 << reg))
+                    MOV(MapReg(reg), W3);
+                else
+                    SaveReg(reg, W3);
+                SetJumpTarget(alreadyWritten);
            }
            else if (!usermode && nextReg != regs.end())
            {
                ARM64Reg first = W3, second = W4;
                
-                if (RegCache.LoadedRegs & (1 << reg) && !(reg == rn && skipLoadingRn))
+                if (RegCache.LoadedRegs & (1 << reg))
                    first = MapReg(reg);
-                if (RegCache.LoadedRegs & (1 << *nextReg) && !(*nextReg == rn && skipLoadingRn))
+                if (RegCache.LoadedRegs & (1 << *nextReg))
                    second = MapReg(*nextReg);

                LDP(INDEX_SIGNED, EncodeRegTo64(first), EncodeRegTo64(second), SP, i * 8);
@ -751,11 +705,8 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
            }
            else if (RegCache.LoadedRegs & (1 << reg))
            {
-                if (!(reg == rn && skipLoadingRn))
-                {
-                    ARM64Reg mapped = MapReg(reg);
-                    LDR(INDEX_UNSIGNED, mapped, SP, i * 8);
-                }
+                ARM64Reg mapped = MapReg(reg);
+                LDR(INDEX_UNSIGNED, mapped, SP, i * 8);
            }
            else
            {
@ -799,13 +750,13 @@ void Compiler::A_Comp_LDM_STM()

    ARM64Reg rn = MapReg(CurInstr.A_Reg(16));

+    s32 offset = Comp_MemAccessBlock(CurInstr.A_Reg(16), regs, !load, pre, !add, usermode);
+
    if (load && writeback && regs[CurInstr.A_Reg(16)])
        writeback = Num == 0
-            && (!(regs & ~BitSet16(1 << CurInstr.A_Reg(16)))) || (regs & ~BitSet16((2 << CurInstr.A_Reg(16)) - 1));
-
-    s32 offset = Comp_MemAccessBlock(CurInstr.A_Reg(16), regs, !load, pre, !add, usermode, load && writeback);
-
-    if (writeback && offset)
+            ? (!(regs & ~BitSet16(1 << CurInstr.A_Reg(16)))) || (regs & ~BitSet16((2 << CurInstr.A_Reg(16)) - 1))
+            : false;
+    if (writeback)
    {
        if (offset > 0)
            ADD(rn, rn, offset);
@ -827,15 +778,12 @@ void Compiler::T_Comp_PUSH_POP()
    }

    ARM64Reg sp = MapReg(13);
-    s32 offset = Comp_MemAccessBlock(13, regs, !load, !load, !load, false, false);
+    s32 offset = Comp_MemAccessBlock(13, regs, !load, !load, !load, false);

-    if (offset)
-    {
-        if (offset > 0)
+    if (offset > 0)
            ADD(sp, sp, offset);
        else
            SUB(sp, sp, -offset);
-    }
 }

 void Compiler::T_Comp_LDMIA_STMIA()
@ -844,12 +792,10 @@ void Compiler::T_Comp_LDMIA_STMIA()
    ARM64Reg rb = MapReg(CurInstr.T_Reg(8));
    bool load = CurInstr.Instr & (1 << 11);
    u32 regsCount = regs.Count();
+    
+    s32 offset = Comp_MemAccessBlock(CurInstr.T_Reg(8), regs, !load, false, false, false);

-    bool writeback = !load || !regs[CurInstr.T_Reg(8)];
-
-    s32 offset = Comp_MemAccessBlock(CurInstr.T_Reg(8), regs, !load, false, false, false, load && writeback);
-
-    if (writeback && offset)
+    if (!load || !regs[CurInstr.T_Reg(8)])
    {
        if (offset > 0)
            ADD(rb, rb, offset);
--- a/src/ARMJIT_Compiler.h
+++ b/src/ARMJIT_Compiler.h
@ -1,26 +1,6 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #ifndef ARMJIT_COMPILER_H
 #define ARMJIT_COMPILER_H

-#ifdef JIT_ENABLED
-
 #if defined(__x86_64__)
 #include "ARMJIT_x64/ARMJIT_Compiler.h"
 #elif defined(__aarch64__)
@ -29,6 +9,9 @@
 #error "The current target platform doesn't have a JIT backend"
 #endif

-#endif
+namespace ARMJIT
+{
+extern Compiler* JITCompiler;
+}

 #endif
--- a/src/ARMJIT_Global.cpp
+++ b/src/ARMJIT_Global.cpp
@ -1,126 +0,0 @@
-#include "ARMJIT_Global.h"
-#include "ARMJIT_Memory.h"
-
-#ifdef _WIN32
-#include <windows.h>
-#else
-#include <sys/mman.h>
-#include <unistd.h>
-#endif
-
-#include <stdio.h>
-#include <stdint.h>
-
-#include <mutex>
-
-namespace melonDS
-{
-
-namespace ARMJIT_Global
-{
-
-std::mutex globalMutex;
-
-#if defined(__APPLE__) && defined(__aarch64__)
-#define APPLE_AARCH64
-#endif
-
-#ifndef APPLE_AARCH64
-static constexpr size_t NumCodeMemSlices = 4;
-static constexpr size_t CodeMemoryAlignedSize = NumCodeMemSlices * CodeMemorySliceSize;
-
-// I haven't heard of pages larger than 16 KB
-u8 CodeMemory[CodeMemoryAlignedSize + 16*1024];
-
-u32 AvailableCodeMemSlices = (1 << NumCodeMemSlices) - 1;
-
-u8* GetAlignedCodeMemoryStart()
-{
-    return reinterpret_cast<u8*>((reinterpret_cast<intptr_t>(CodeMemory) + (16*1024-1)) & ~static_cast<intptr_t>(16*1024-1));
-}
-#endif
-
-int RefCounter = 0;
-
-void* AllocateCodeMem()
-{
-    std::lock_guard guard(globalMutex);
-
-#ifndef APPLE_AARCH64
-    if (AvailableCodeMemSlices)
-    {
-        int slice = __builtin_ctz(AvailableCodeMemSlices);
-        AvailableCodeMemSlices &= ~(1 << slice);
-        //printf("allocating slice %d\n", slice);
-        return &GetAlignedCodeMemoryStart()[slice * CodeMemorySliceSize];
-    }
-#endif
-
-    // allocate
-#ifdef _WIN32
-    return VirtualAlloc(nullptr, CodeMemorySliceSize, MEM_RESERVE|MEM_COMMIT, PAGE_EXECUTE_READWRITE);
-#elif defined(APPLE_AARCH64)
-    return mmap(NULL, CodeMemorySliceSize, PROT_READ | PROT_WRITE | PROT_EXEC, MAP_PRIVATE | MAP_ANONYMOUS | MAP_JIT,-1, 0);
-#else
-    //printf("mmaping...\n");
-    return mmap(nullptr, CodeMemorySliceSize, PROT_READ | PROT_WRITE | PROT_EXEC, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
-#endif
-}
-
-void FreeCodeMem(void* codeMem)
-{
-    std::lock_guard guard(globalMutex);
-
-#ifndef APPLE_AARCH64
-    for (int i = 0; i < NumCodeMemSlices; i++)
-    {
-        if (codeMem == &GetAlignedCodeMemoryStart()[CodeMemorySliceSize * i])
-        {
-            //printf("freeing slice\n");
-            AvailableCodeMemSlices |= 1 << i;
-            return;
-        }
-    }
-#endif
-
-#ifdef _WIN32
-    VirtualFree(codeMem, CodeMemorySliceSize, MEM_RELEASE|MEM_DECOMMIT);
-#else
-    munmap(codeMem, CodeMemorySliceSize);
-#endif
-}
-
-void Init()
-{
-    std::lock_guard guard(globalMutex);
-
-    RefCounter++;
-    if (RefCounter == 1)
-    {
-        #ifdef _WIN32
-            DWORD dummy;
-            VirtualProtect(GetAlignedCodeMemoryStart(), CodeMemoryAlignedSize, PAGE_EXECUTE_READWRITE, &dummy);
-        #elif defined(APPLE_AARCH64)
-            // Apple aarch64 always uses dynamic allocation
-        #else
-            mprotect(GetAlignedCodeMemoryStart(), CodeMemoryAlignedSize, PROT_EXEC | PROT_READ | PROT_WRITE);
-        #endif
-
-        ARMJIT_Memory::RegisterFaultHandler();
-    }
-}
-
-void DeInit()
-{
-    std::lock_guard guard(globalMutex);
-
-    RefCounter--;
-    if (RefCounter == 0)
-    {
-        ARMJIT_Memory::UnregisterFaultHandler();
-    }
-}
-
-}
-
-}
--- a/src/ARMJIT_Internal.h
+++ b/src/ARMJIT_Internal.h
@ -1,21 +1,3 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #ifndef ARMJIT_INTERNAL_H
 #define ARMJIT_INTERNAL_H

@ -24,17 +6,13 @@
 #include <string.h>
 #include <assert.h>

-#include "ARM_InstrInfo.h"
-#include "JitBlock.h"
-#include "TinyVector.h"
-
-namespace melonDS
-{
-class ARM;
-class ARMv5;
+#include "ARMJIT.h"
+#include "ARMJIT_Memory.h"

 // here lands everything which doesn't fit into ARMJIT.h
 // where it would be included by pretty much everything
+namespace ARMJIT
+{

 enum
 {
@ -73,6 +51,139 @@ struct FetchedInstr
    ARMInstrInfo::Info Info;
 };

+/*
+    TinyVector
+        - because reinventing the wheel is the best!
+    
+    - meant to be used very often, with not so many elements
+    max 1 << 16 elements
+    - doesn't allocate while no elements are inserted
+    - not stl confirmant of course
+    - probably only works with POD types
+    - remove operations don't preserve order, but O(1)!
+*/
+template <typename T>
+struct __attribute__((packed)) TinyVector
+{
+    T* Data = NULL;
+    u16 Capacity = 0;
+    u16 Length = 0;
+
+    ~TinyVector()
+    {
+        delete[] Data;
+    }
+
+    void MakeCapacity(u32 capacity)
+    {
+        assert(capacity <= UINT16_MAX);
+        assert(capacity > Capacity);
+        T* newMem = new T[capacity];
+        if (Data != NULL)
+            memcpy(newMem, Data, sizeof(T) * Length);
+
+        T* oldData = Data;
+        Data = newMem;
+        if (oldData != NULL)
+            delete[] oldData;
+        
+        Capacity = capacity;
+    }
+
+    void SetLength(u16 length)
+    {
+        if (Capacity < length)
+            MakeCapacity(length);
+        
+        Length = length;
+    }
+
+    void Clear()
+    {
+        Length = 0;
+    }
+
+    void Add(T element)
+    {
+        assert(Length + 1 <= UINT16_MAX);
+        if (Length + 1 > Capacity)
+            MakeCapacity(((Capacity + 4) * 3) / 2);
+        
+        Data[Length++] = element;
+    }
+
+    void Remove(int index)
+    {
+        assert(Length > 0);
+        assert(index >= 0 && index < Length);
+
+        Length--;
+        Data[index] = Data[Length];
+        /*for (int i = index; i < Length; i++)
+            Data[i] = Data[i + 1];*/
+    }
+
+    int Find(T needle)
+    {
+        for (int i = 0; i < Length; i++)
+        {
+            if (Data[i] == needle)
+                return i;
+        }
+        return -1;
+    }
+
+    bool RemoveByValue(T needle)
+    {
+        for (int i = 0; i < Length; i++)
+        {
+            if (Data[i] == needle)
+            {
+                Remove(i);
+                return true;
+            }
+        }
+        return false;
+    }
+
+    T& operator[](int index)
+    {
+        assert(index >= 0 && index < Length);
+        return Data[index];
+    }
+};
+
+class JitBlock
+{
+public:
+    JitBlock(u32 num, u32 literalHash, u32 numAddresses, u32 numLiterals)
+    {
+        Num = num;
+        NumAddresses = numAddresses;
+        NumLiterals = numLiterals;
+        Data.SetLength(numAddresses * 2 + numLiterals);
+    }
+
+    u32 StartAddr;
+    u32 StartAddrLocal;
+    u32 InstrHash, LiteralHash;
+    u8 Num;
+    u16 NumAddresses;
+    u16 NumLiterals;
+
+    JitBlockEntry EntryPoint;
+
+    u32* AddressRanges()
+    { return &Data[0]; }
+    u32* AddressMasks()
+    { return &Data[NumAddresses]; }
+    u32* Literals()
+    { return &Data[NumAddresses * 2]; }
+
+private:
+    TinyVector<u32> Data;
+};
+
 // size should be 16 bytes because I'm to lazy to use mul and whatnot
 struct __attribute__((packed)) AddressRange
 {
@ -85,9 +196,13 @@ typedef void (*InterpreterFunc)(ARM* cpu);
 extern InterpreterFunc InterpretARM[];
 extern InterpreterFunc InterpretTHUMB[];

-inline bool PageContainsCode(const AddressRange* range, u32 pageSize)
+extern TinyVector<u32> InvalidLiterals;
+
+extern AddressRange* const CodeMemRegions[ARMJIT_Memory::memregions_Count];
+
+inline bool PageContainsCode(AddressRange* range)
 {
-    for (int i = 0; i < pageSize / 512; i++)
+    for (int i = 0; i < 8; i++)
    {
        if (range[i].Blocks.Length > 0)
            return true;
@ -95,6 +210,11 @@ inline bool PageContainsCode(const AddressRange* range, u32 pageSize)
    return false;
 }

+u32 LocaliseCodeAddress(u32 num, u32 addr);
+
+template <u32 Num>
+void LinkBlock(ARM* cpu, u32 codeOffset);
+
 template <typename T, int ConsoleType> T SlowRead9(u32 addr, ARMv5* cpu);
 template <typename T, int ConsoleType> void SlowWrite9(u32 addr, ARMv5* cpu, u32 val);
 template <typename T, int ConsoleType> T SlowRead7(u32 addr);
--- a/src/ARMJIT_Memory.cpp
+++ b/src/ARMJIT_Memory.cpp
--- a/src/ARMJIT_Memory.h
+++ b/src/ARMJIT_Memory.h
@ -1,244 +1,63 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #ifndef ARMJIT_MEMORY
 #define ARMJIT_MEMORY

 #include "types.h"
-#include "MemConstants.h"

-#ifdef JIT_ENABLED
-#  include <mutex>
-#  include "TinyVector.h"
-#  include "ARM.h"
-#  if defined(__SWITCH__)
-#    include <switch.h>
-#  elif defined(_WIN32)
-#include <windows.h>
-#  else
-#    include <sys/mman.h>
-#    include <sys/stat.h>
-#    include <fcntl.h>
-#    include <unistd.h>
-#    include <signal.h>
-#  endif
-#else
-# include <array>
-#endif
+#include "ARM.h"

-namespace melonDS
+namespace ARMJIT_Memory
 {
-#ifdef JIT_ENABLED
-namespace Platform { struct DynamicLibrary; }
-class Compiler;
-class ARMJIT;
-#endif

-static constexpr u32 LargePageSize = 0x4000;
-static constexpr u32 RegularPageSize = 0x1000;
+extern void* FastMem9Start;
+extern void* FastMem7Start;

-constexpr u32 RoundUp(u32 size) noexcept
+void Init();
+void DeInit();
+
+void Reset();
+
+enum
 {
-    return (size + LargePageSize - 1) & ~(LargePageSize - 1);
-}
+    memregion_Other = 0,
+    memregion_ITCM,
+    memregion_DTCM,
+    memregion_BIOS9,
+    memregion_MainRAM,
+    memregion_SharedWRAM,
+    memregion_IO9,
+    memregion_VRAM,
+    memregion_BIOS7,
+    memregion_WRAM7,
+    memregion_IO7,
+    memregion_Wifi,
+    memregion_VWRAM,

-static constexpr u32 MemBlockMainRAMOffset = 0;
-static constexpr u32 MemBlockSWRAMOffset = RoundUp(MainRAMMaxSize);
-static constexpr u32 MemBlockARM7WRAMOffset = MemBlockSWRAMOffset + RoundUp(SharedWRAMSize);
-static constexpr u32 MemBlockDTCMOffset = MemBlockARM7WRAMOffset + RoundUp(ARM7WRAMSize);
-static constexpr u32 MemBlockNWRAM_AOffset = MemBlockDTCMOffset + RoundUp(DTCMPhysicalSize);
-static constexpr u32 MemBlockNWRAM_BOffset = MemBlockNWRAM_AOffset + RoundUp(NWRAMSize);
-static constexpr u32 MemBlockNWRAM_COffset = MemBlockNWRAM_BOffset + RoundUp(NWRAMSize);
-static constexpr u32 MemoryTotalSize = MemBlockNWRAM_COffset + RoundUp(NWRAMSize);
+    // DSi
+    memregion_BIOS9DSi,
+    memregion_BIOS7DSi,
+    memregion_NewSharedWRAM_A,
+    memregion_NewSharedWRAM_B,
+    memregion_NewSharedWRAM_C,

-class ARMJIT_Memory
-{
-public:
-    enum
-    {
-        memregion_Other = 0,
-        memregion_ITCM,
-        memregion_DTCM,
-        memregion_BIOS9,
-        memregion_MainRAM,
-        memregion_SharedWRAM,
-        memregion_IO9,
-        memregion_VRAM,
-        memregion_BIOS7,
-        memregion_WRAM7,
-        memregion_IO7,
-        memregion_Wifi,
-        memregion_VWRAM,
-
-        // DSi
-        memregion_BIOS9DSi,
-        memregion_BIOS7DSi,
-        memregion_NewSharedWRAM_A,
-        memregion_NewSharedWRAM_B,
-        memregion_NewSharedWRAM_C,
-
-        memregions_Count
-    };
-
-#ifdef JIT_ENABLED
-public:
-    explicit ARMJIT_Memory(melonDS::NDS& nds);
-    ~ARMJIT_Memory() noexcept;
-    ARMJIT_Memory(const ARMJIT_Memory&) = delete;
-    ARMJIT_Memory(ARMJIT_Memory&&) = delete;
-    ARMJIT_Memory& operator=(const ARMJIT_Memory&) = delete;
-    ARMJIT_Memory& operator=(ARMJIT_Memory&&) = delete;
-    void Reset() noexcept;
-    void RemapDTCM(u32 newBase, u32 newSize) noexcept;
-    void RemapSWRAM() noexcept;
-    void RemapNWRAM(int num) noexcept;
-    void SetCodeProtection(int region, u32 offset, bool protect) noexcept;
-
-    [[nodiscard]] u8* GetMainRAM() noexcept { return MemoryBase + MemBlockMainRAMOffset; }
-    [[nodiscard]] const u8* GetMainRAM() const noexcept { return MemoryBase + MemBlockMainRAMOffset; }
-
-    [[nodiscard]] u8* GetSharedWRAM() noexcept { return MemoryBase + MemBlockSWRAMOffset; }
-    [[nodiscard]] const u8* GetSharedWRAM() const noexcept { return MemoryBase + MemBlockSWRAMOffset; }
-
-    [[nodiscard]] u8* GetARM7WRAM() noexcept { return MemoryBase + MemBlockARM7WRAMOffset; }
-    [[nodiscard]] const u8* GetARM7WRAM() const noexcept { return MemoryBase + MemBlockARM7WRAMOffset; }
-
-    [[nodiscard]] u8* GetARM9DTCM() noexcept { return MemoryBase + MemBlockDTCMOffset; }
-    [[nodiscard]] const u8* GetARM9DTCM() const noexcept { return MemoryBase + MemBlockDTCMOffset; }
-
-    [[nodiscard]] u8* GetNWRAM_A() noexcept { return MemoryBase + MemBlockNWRAM_AOffset; }
-    [[nodiscard]] const u8* GetNWRAM_A() const noexcept { return MemoryBase + MemBlockNWRAM_AOffset; }
-
-    [[nodiscard]] u8* GetNWRAM_B() noexcept { return MemoryBase + MemBlockNWRAM_BOffset; }
-    [[nodiscard]] const u8* GetNWRAM_B() const noexcept { return MemoryBase + MemBlockNWRAM_BOffset; }
-
-    [[nodiscard]] u8* GetNWRAM_C() noexcept { return MemoryBase + MemBlockNWRAM_COffset; }
-    [[nodiscard]] const u8* GetNWRAM_C() const noexcept { return MemoryBase + MemBlockNWRAM_COffset; }
-
-    int ClassifyAddress9(u32 addr) const noexcept;
-    int ClassifyAddress7(u32 addr) const noexcept;
-    bool GetMirrorLocation(int region, u32 num, u32 addr, u32& memoryOffset, u32& mirrorStart, u32& mirrorSize) const noexcept;
-    u32 LocaliseAddress(int region, u32 num, u32 addr) const noexcept;
-    bool IsFastmemCompatible(int region) const noexcept;
-    void* GetFuncForAddr(ARM* cpu, u32 addr, bool store, int size) const noexcept;
-    bool MapAtAddress(u32 addr) noexcept;
-
-    static bool IsFastMemSupported();
-
-    static void RegisterFaultHandler();
-    static void UnregisterFaultHandler();
-
-    static u32 PageSize;
-    static u32 PageShift;
-private:
-    friend class Compiler;
-    struct Mapping
-    {
-        u32 Addr;
-        u32 Size, LocalOffset;
-        u32 Num;
-
-        void Unmap(int region, NDS& nds) noexcept;
-    };
-
-    struct FaultDescription
-    {
-        u32 EmulatedFaultAddr;
-        u8* FaultPC;
-    };
-    static bool FaultHandler(FaultDescription& faultDesc, melonDS::NDS& nds);
-    bool MapIntoRange(u32 addr, u32 num, u32 offset, u32 size) noexcept;
-    bool UnmapFromRange(u32 addr, u32 num, u32 offset, u32 size) noexcept;
-    void SetCodeProtectionRange(u32 addr, u32 size, u32 num, int protection) noexcept;
-
-    melonDS::NDS& NDS;
-    void* FastMem9Start;
-    void* FastMem7Start;
-    u8* MemoryBase = nullptr;
-
-#if defined(__SWITCH__)
-    VirtmemReservation* FastMem9Reservation, *FastMem7Reservation;
-    u8* MemoryBaseCodeMem;
-#elif defined(_WIN32)
-    struct VirtmemPlaceholder
-    {
-        uintptr_t Start;
-        size_t Size;
-    };
-    std::vector<VirtmemPlaceholder> VirtmemPlaceholders;
-
-    static LONG ExceptionHandler(EXCEPTION_POINTERS* exceptionInfo);
-    HANDLE MemoryFile = INVALID_HANDLE_VALUE;
-#else
-
-    static void SigsegvHandler(int sig, siginfo_t* info, void* rawContext);
-    int MemoryFile = -1;
-#endif
-#ifdef ANDROID
-    Platform::DynamicLibrary* Libandroid = nullptr;
-#endif
-    u8 MappingStatus9[1 << (32-12)] {};
-    u8 MappingStatus7[1 << (32-12)] {};
-    TinyVector<Mapping> Mappings[memregions_Count] {};
-#else
-public:
-    explicit ARMJIT_Memory(melonDS::NDS&) {};
-    ~ARMJIT_Memory() = default;
-    ARMJIT_Memory(const ARMJIT_Memory&) = delete;
-    ARMJIT_Memory(ARMJIT_Memory&&) = delete;
-    ARMJIT_Memory& operator=(const ARMJIT_Memory&) = delete;
-    ARMJIT_Memory& operator=(ARMJIT_Memory&&) = delete;
-
-    void Reset() noexcept {}
-    void RemapDTCM(u32 newBase, u32 newSize) noexcept {}
-    void RemapSWRAM() noexcept {}
-    void RemapNWRAM(int num) noexcept {}
-    void SetCodeProtection(int region, u32 offset, bool protect) noexcept {}
-
-    [[nodiscard]] u8* GetMainRAM() noexcept { return MainRAM.data(); }
-    [[nodiscard]] const u8* GetMainRAM() const noexcept { return MainRAM.data(); }
-
-    [[nodiscard]] u8* GetSharedWRAM() noexcept { return SharedWRAM.data(); }
-    [[nodiscard]] const u8* GetSharedWRAM() const noexcept { return SharedWRAM.data(); }
-
-    [[nodiscard]] u8* GetARM7WRAM() noexcept { return ARM7WRAM.data(); }
-    [[nodiscard]] const u8* GetARM7WRAM() const noexcept { return ARM7WRAM.data(); }
-
-    [[nodiscard]] u8* GetARM9DTCM() noexcept { return DTCM.data(); }
-    [[nodiscard]] const u8* GetARM9DTCM() const noexcept { return DTCM.data(); }
-
-    [[nodiscard]] u8* GetNWRAM_A() noexcept { return NWRAM_A.data(); }
-    [[nodiscard]] const u8* GetNWRAM_A() const noexcept { return NWRAM_A.data(); }
-
-    [[nodiscard]] u8* GetNWRAM_B() noexcept { return NWRAM_B.data(); }
-    [[nodiscard]] const u8* GetNWRAM_B() const noexcept { return NWRAM_B.data(); }
-
-    [[nodiscard]] u8* GetNWRAM_C() noexcept { return NWRAM_C.data(); }
-    [[nodiscard]] const u8* GetNWRAM_C() const noexcept { return NWRAM_C.data(); }
-private:
-    std::array<u8, MainRAMMaxSize> MainRAM {};
-    std::array<u8, ARM7WRAMSize> ARM7WRAM {};
-    std::array<u8, SharedWRAMSize> SharedWRAM {};
-    std::array<u8, DTCMPhysicalSize> DTCM {};
-    std::array<u8, NWRAMSize> NWRAM_A {};
-    std::array<u8, NWRAMSize> NWRAM_B {};
-    std::array<u8, NWRAMSize> NWRAM_C {};
-#endif
+    memregions_Count
 };
+
+int ClassifyAddress9(u32 addr);
+int ClassifyAddress7(u32 addr);
+
+bool GetMirrorLocation(int region, u32 num, u32 addr, u32& memoryOffset, u32& mirrorStart, u32& mirrorSize);
+u32 LocaliseAddress(int region, u32 num, u32 addr);
+
+bool IsFastmemCompatible(int region);
+
+void RemapDTCM(u32 newBase, u32 newSize);
+void RemapSWRAM();
+void RemapNWRAM(int num);
+
+void SetCodeProtection(int region, u32 offset, bool protect);
+
+void* GetFuncForAddr(ARM* cpu, u32 addr, bool store, int size);
+
 }
-#endif
+
+#endif
--- a/src/ARMJIT_RegisterCache.h
+++ b/src/ARMJIT_RegisterCache.h
@ -1,38 +1,15 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #ifndef ARMJIT_REGCACHE_H
 #define ARMJIT_REGCACHE_H

-#include "ARMJIT_Internal.h"
-#include "Platform.h"
+#include "ARMJIT.h"

 // TODO: replace this in the future
 #include "dolphin/BitSet.h"

 #include <assert.h>

-namespace melonDS
+namespace ARMJIT
 {
-    using Platform::Log;
-    using Platform::LogLevel;
-    using namespace Common;
-    // Imported inside the namespace so that other headers aren't polluted

 template <typename T, typename Reg>
 class RegisterCache
@ -46,7 +23,7 @@ public:
    {
        for (int i = 0; i < 16; i++)
            Mapping[i] = (Reg)-1;
-
+        
        PCAllocatableAsSrc = ~(pcAllocatableAsSrc
            ? 0
            : (1 << 15));
@ -84,7 +61,7 @@ public:
            }
        }

-        Log(LogLevel::Error, "this is a JIT bug! LoadRegister failed\n");
+        printf("this is a JIT bug! LoadRegister failed\n");
        abort();
    }

@ -99,7 +76,7 @@ public:
        LiteralsLoaded &= ~(1 << reg);
    }

-    bool IsLiteral(int reg) const
+    bool IsLiteral(int reg)
    {
        return LiteralsLoaded & (1 << reg);
    }
--- a/src/ARMJIT_x64/ARMJIT_ALU.cpp
+++ b/src/ARMJIT_x64/ARMJIT_ALU.cpp
@ -1,27 +1,8 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARMJIT_Compiler.h"
-#include "../ARM.h"

 using namespace Gen;

-namespace melonDS
+namespace ARMJIT
 {

 // uses RSCRATCH3
@ -129,7 +110,7 @@ OpArg Compiler::A_Comp_GetALUOp2(bool S, bool& carryUsed)
        Comp_AddCycles_C();

        u32 shift = (CurInstr.Instr >> 7) & 0x1E;
-        u32 imm = melonDS::ROR(CurInstr.Instr & 0xFF, shift);
+        u32 imm = ::ROR(CurInstr.Instr & 0xFF, shift);

        carryUsed = false;
        if (S && shift)
@ -228,7 +209,7 @@ void Compiler::A_Comp_Arith()
        Comp_ArithTriOp(&Compiler::AND, rd, rn, op2, carryUsed, sFlag|opSymmetric|opInvertOp2);
        break;
    default:
-        Log(LogLevel::Error, "this is a JIT bug! %04x\n", op);
+        printf("this is a JIT bug! %04x\n", op);
        abort();
    }

@ -365,7 +346,7 @@ void Compiler::A_Comp_Mul_Long()
        {
            BSR(32, RSCRATCH, R(RSCRATCH3));
        }
-
+        
        SHR(32, R(RSCRATCH), Imm8(3));
        SetJumpTarget(zeroBSR); // fortunately that's even right
        Comp_AddCycles_CI(RSCRATCH, 2);
@ -618,7 +599,7 @@ void Compiler::T_Comp_AddSub_()
    int op = (CurInstr.Instr >> 9) & 0x3;

    OpArg rn = op >= 2 ? Imm32((CurInstr.Instr >> 6) & 0x7) : MapReg(CurInstr.T_Reg(6));
-
+    
    Comp_AddCycles_C();

    // special case for thumb mov being alias to add rd, rn, #0
--- a/src/ARMJIT_x64/ARMJIT_Branch.cpp
+++ b/src/ARMJIT_x64/ARMJIT_Branch.cpp
@ -1,28 +1,8 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARMJIT_Compiler.h"
-#include "../ARM.h"
-#include "../NDS.h"

 using namespace Gen;

-namespace melonDS
+namespace ARMJIT
 {

 template <typename T>
@ -32,7 +12,7 @@ int squeezePointer(T* ptr)
    assert((T*)((u64)truncated) == ptr);
    return truncated;
 }
-
+    
 void Compiler::Comp_JumpTo(u32 addr, bool forceNonConstantCycles)
 {
    // we can simplify constant branches by a lot
@ -121,7 +101,7 @@ void Compiler::Comp_JumpTo(u32 addr, bool forceNonConstantCycles)
            u32 compileTimePC = CurCPU->R[15];
            CurCPU->R[15] = newPC;

-            cycles += NDS.ARM7MemTimings[codeCycles][0] + NDS.ARM7MemTimings[codeCycles][1];
+            cycles += NDS::ARM7MemTimings[codeCycles][0] + NDS::ARM7MemTimings[codeCycles][1];

            CurCPU->R[15] = compileTimePC;
        }
@ -133,7 +113,7 @@ void Compiler::Comp_JumpTo(u32 addr, bool forceNonConstantCycles)
            u32 compileTimePC = CurCPU->R[15];
            CurCPU->R[15] = newPC;

-            cycles += NDS.ARM7MemTimings[codeCycles][2] + NDS.ARM7MemTimings[codeCycles][3];
+            cycles += NDS::ARM7MemTimings[codeCycles][2] + NDS::ARM7MemTimings[codeCycles][3];

            CurCPU->R[15] = compileTimePC;
        }
@ -167,7 +147,7 @@ void Compiler::Comp_JumpTo(Gen::X64Reg addr, bool restoreCPSR)
    bool cpsrDirty = CPSRDirty;
    SaveCPSR();

-    PushRegs(restoreCPSR, true);
+    PushRegs(restoreCPSR);

    MOV(64, R(ABI_PARAM1), R(RCPU));
    MOV(32, R(ABI_PARAM2), R(addr));
@ -176,11 +156,11 @@ void Compiler::Comp_JumpTo(Gen::X64Reg addr, bool restoreCPSR)
    else
        MOV(32, R(ABI_PARAM3), Imm32(true)); // what a waste
    if (Num == 0)
-        ABI_CallFunction(ARMv5JumpToTrampoline);
+        CALL((void*)&ARMv5JumpToTrampoline);
    else
-        ABI_CallFunction(ARMv4JumpToTrampoline);
+        CALL((void*)&ARMv4JumpToTrampoline);

-    PopRegs(restoreCPSR, true);
+    PopRegs(restoreCPSR);

    LoadCPSR();
    // in case this instruction is skipped
@ -249,7 +229,7 @@ void Compiler::T_Comp_BranchXchangeReg()
    {
        if (Num == 1)
        {
-            Log(LogLevel::Warn, "BLX unsupported on ARM7!!!\n");
+            printf("BLX unsupported on ARM7!!!\n");
            return;
        }
        MOV(32, R(RSCRATCH), MapReg(CurInstr.A_Reg(3)));
@ -295,7 +275,7 @@ void Compiler::T_Comp_BL_Merged()
        target |= 1;

    MOV(32, MapReg(14), Imm32((R15 - 2) | 1));
-
+    
    Comp_JumpTo(target);
 }

--- a/src/ARMJIT_x64/ARMJIT_Compiler.cpp
+++ b/src/ARMJIT_x64/ARMJIT_Compiler.cpp
@ -1,39 +1,24 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARMJIT_Compiler.h"

-#include "../ARMJIT.h"
 #include "../ARMInterpreter.h"
-#include "../NDS.h"
-#include "../ARMJIT_Global.h"
+#include "../Config.h"

 #include <assert.h>
-#include <stdarg.h>

 #include "../dolphin/CommonFuncs.h"

+#ifdef _WIN32
+#include <windows.h>
+#else
+#include <sys/mman.h>
+#include <unistd.h>
+#endif
+
 using namespace Gen;
-using namespace Common;

 extern "C" void ARM_Ret();

-namespace melonDS
+namespace ARMJIT
 {
 template <>
 const X64Reg RegisterCache<Compiler, X64Reg>::NativeRegAllocOrder[] =
@ -61,7 +46,7 @@ const BitSet32 CallerSavedPushRegs({R10, R11});
 const BitSet32 CallerSavedPushRegs({R9, R10, R11});
 #endif

-void Compiler::PushRegs(bool saveHiRegs, bool saveRegsToBeChanged, bool allowUnload)
+void Compiler::PushRegs(bool saveHiRegs)
 {
    BitSet32 loadedRegs(RegCache.LoadedRegs);

@ -80,26 +65,17 @@ void Compiler::PushRegs(bool saveHiRegs, bool saveRegsToBeChanged, bool allowUnl
    }

    for (int reg : loadedRegs)
-    {
-        if (CallerSavedPushRegs[RegCache.Mapping[reg]]
-            && (saveRegsToBeChanged || !((1<<reg) & CurInstr.Info.DstRegs && !((1<<reg) & CurInstr.Info.SrcRegs))))
-        {
-            if ((Thumb || CurInstr.Cond() == 0xE) && !((1 << reg) & (CurInstr.Info.DstRegs|CurInstr.Info.SrcRegs)) && allowUnload)
-                RegCache.UnloadRegister(reg);
-            else
-                SaveReg(reg, RegCache.Mapping[reg]);
-        }
-    }
+        if (BitSet32(1 << RegCache.Mapping[reg]) & ABI_ALL_CALLER_SAVED)
+            SaveReg(reg, RegCache.Mapping[reg]);
 }

-void Compiler::PopRegs(bool saveHiRegs, bool saveRegsToBeChanged)
+void Compiler::PopRegs(bool saveHiRegs)
 {
    BitSet32 loadedRegs(RegCache.LoadedRegs);
    for (int reg : loadedRegs)
    {
        if ((saveHiRegs && reg >= 8 && reg < 15)
-            || (CallerSavedPushRegs[RegCache.Mapping[reg]]
-                && (saveRegsToBeChanged || !((1<<reg) & CurInstr.Info.DstRegs && !((1<<reg) & CurInstr.Info.SrcRegs)))))
+            || BitSet32(1 << RegCache.Mapping[reg]) & ABI_ALL_CALLER_SAVED)
        {
            LoadReg(reg, RegCache.Mapping[reg]);
        }
@ -122,9 +98,7 @@ void Compiler::A_Comp_MRS()
        MOV(32, rd, R(RSCRATCH3));
    }
    else
-    {
        MOV(32, rd, R(RCPSR));
-    }
 }

 void UpdateModeTrampoline(ARM* arm, u32 oldmode, u32 newmode)
@ -137,7 +111,7 @@ void Compiler::A_Comp_MSR()
    Comp_AddCycles_C();

    OpArg val = CurInstr.Instr & (1 << 25)
-        ? Imm32(melonDS::ROR((CurInstr.Instr & 0xFF), ((CurInstr.Instr >> 7) & 0x1E)))
+        ? Imm32(::ROR((CurInstr.Instr & 0xFF), ((CurInstr.Instr >> 7) & 0x1E)))
        : MapReg(CurInstr.A_Reg(0));

    u32 mask = 0;
@ -211,26 +185,51 @@ void Compiler::A_Comp_MSR()
            AND(32, R(RSCRATCH2), val);
            OR(32, R(RCPSR), R(RSCRATCH2));

-            PushRegs(true, true);
+            PushRegs(true);

            MOV(32, R(ABI_PARAM3), R(RCPSR));
            MOV(32, R(ABI_PARAM2), R(RSCRATCH3));
            MOV(64, R(ABI_PARAM1), R(RCPU));
-            ABI_CallFunction(UpdateModeTrampoline);
+            CALL((void*)&UpdateModeTrampoline);

-            PopRegs(true, true);
+            PopRegs(true);
        }
    }
 }

-Compiler::Compiler(melonDS::NDS& nds) : XEmitter(), NDS(nds)
+/*
+    We'll repurpose this .bss memory
+
+ */
+u8 CodeMemory[1024 * 1024 * 32];
+
+Compiler::Compiler()
 {
-    ARMJIT_Global::Init();
+    {
+    #ifdef _WIN32
+        SYSTEM_INFO sysInfo;
+        GetSystemInfo(&sysInfo);

-    CodeMemBase = static_cast<u8*>(ARMJIT_Global::AllocateCodeMem());
-    CodeMemSize = ARMJIT_Global::CodeMemorySliceSize;
+        u64 pageSize = (u64)sysInfo.dwPageSize;
+    #else
+        u64 pageSize = sysconf(_SC_PAGE_SIZE);
+    #endif

-    ResetStart = CodeMemBase;
+        u8* pageAligned = (u8*)(((u64)CodeMemory & ~(pageSize - 1)) + pageSize);
+        u64 alignedSize = (((u64)CodeMemory + sizeof(CodeMemory)) & ~(pageSize - 1)) - (u64)pageAligned;
+
+    #ifdef _WIN32
+        DWORD dummy;
+        VirtualProtect(pageAligned, alignedSize, PAGE_EXECUTE_READWRITE, &dummy);
+    #elif defined(__APPLE__)
+        pageAligned = (u8*)mmap(NULL, 1024*1024*32, PROT_READ | PROT_WRITE | PROT_EXEC, MAP_PRIVATE | MAP_ANONYMOUS ,-1, 0);
+    #else
+        mprotect(pageAligned, alignedSize, PROT_EXEC | PROT_READ | PROT_WRITE);
+    #endif
+
+        ResetStart = pageAligned;
+        CodeMemSize = alignedSize;
+    }

    Reset();

@ -270,10 +269,6 @@ Compiler::Compiler(melonDS::NDS& nds) : XEmitter(), NDS(nds)
        SetJumpTarget(und);
        MOV(32, R(RSCRATCH3), MComplex(RCPU, RSCRATCH2, SCALE_4, offsetof(ARM, R_UND)));
        RET();
-
-#ifdef JIT_PROFILING_ENABLED
-        CreateMethod("ReadBanked", ReadBanked);
-#endif
    }
    {
        // RSCRATCH  mode
@ -317,10 +312,6 @@ Compiler::Compiler(melonDS::NDS& nds) : XEmitter(), NDS(nds)
        MOV(32, MComplex(RCPU, RSCRATCH2, SCALE_4, offsetof(ARM, R_UND)), R(RSCRATCH3));
        CLC();
        RET();
-
-#ifdef JIT_PROFILING_ENABLED
-        CreateMethod("WriteBanked", WriteBanked);
-#endif
    }

    for (int consoleType = 0; consoleType < 2; consoleType++)
@ -381,10 +372,6 @@ Compiler::Compiler(melonDS::NDS& nds) : XEmitter(), NDS(nds)
                    ABI_PopRegistersAndAdjustStack(CallerSavedPushRegs, 8);
                    RET();

-#ifdef JIT_PROFILING_ENABLED
-                    CreateMethod("FastMemStorePatch%d_%d_%d", PatchedStoreFuncs[consoleType][num][size][reg], num, size, reg);
-#endif
-
                    for (int signextend = 0; signextend < 2; signextend++)
                    {
                        PatchedLoadFuncs[consoleType][num][size][signextend][reg] = GetWritableCodePtr();
@ -423,10 +410,6 @@ Compiler::Compiler(melonDS::NDS& nds) : XEmitter(), NDS(nds)
                        else
                            MOVZX(32, 8 << size, rdMapped, R(RSCRATCH));
                        RET();
-
-#ifdef JIT_PROFILING_ENABLED
-                        CreateMethod("FastMemLoadPatch%d_%d_%d_%d", PatchedLoadFuncs[consoleType][num][size][signextend][reg], num, size, reg, signextend);
-#endif
                    }
                }
            }
@ -444,13 +427,6 @@ Compiler::Compiler(melonDS::NDS& nds) : XEmitter(), NDS(nds)
    FarSize = (ResetStart + CodeMemSize) - FarStart;
 }

-Compiler::~Compiler()
-{
-    ARMJIT_Global::FreeCodeMem(CodeMemBase);
-
-    ARMJIT_Global::DeInit();
-}
-
 void Compiler::LoadCPSR()
 {
    assert(!CPSRDirty);
@ -471,14 +447,14 @@ void Compiler::SaveCPSR(bool flagClean)
 void Compiler::LoadReg(int reg, X64Reg nativeReg)
 {
    if (reg != 15)
-        MOV(32, R(nativeReg), MDisp(RCPU, offsetof(ARM, R) + reg*4));
+        MOV(32, R(nativeReg), MDisp(RCPU, offsetof(ARM, R[reg])));
    else
        MOV(32, R(nativeReg), Imm32(R15));
 }

 void Compiler::SaveReg(int reg, X64Reg nativeReg)
 {
-    MOV(32, MDisp(RCPU, offsetof(ARM, R) + reg*4), R(nativeReg));
+    MOV(32, MDisp(RCPU, offsetof(ARM, R[reg])), R(nativeReg));
 }

 // invalidates RSCRATCH and RSCRATCH3
@ -616,18 +592,18 @@ const Compiler::CompileFunc T_Comp[ARMInstrInfo::tk_Count] = {
    // LDR/STR sp rel
    F(T_Comp_MemSPRel), F(T_Comp_MemSPRel),
    // PUSH/POP
-    F(T_Comp_PUSH_POP), F(T_Comp_PUSH_POP),
+    F(T_Comp_PUSH_POP), F(T_Comp_PUSH_POP), 
    // LDMIA, STMIA
-    F(T_Comp_LDMIA_STMIA), F(T_Comp_LDMIA_STMIA),
+    F(T_Comp_LDMIA_STMIA), F(T_Comp_LDMIA_STMIA), 
    // Branch
-    F(T_Comp_BCOND), F(T_Comp_BranchXchangeReg), F(T_Comp_BranchXchangeReg), F(T_Comp_B), F(T_Comp_BL_LONG_1), F(T_Comp_BL_LONG_2),
+    F(T_Comp_BCOND), F(T_Comp_BranchXchangeReg), F(T_Comp_BranchXchangeReg), F(T_Comp_B), F(T_Comp_BL_LONG_1), F(T_Comp_BL_LONG_2), 
    // Unk, SVC
    NULL, NULL,
    F(T_Comp_BL_Merged)
 };
 #undef F

-bool Compiler::CanCompile(bool thumb, u16 kind) const
+bool Compiler::CanCompile(bool thumb, u16 kind)
 {
    return (thumb ? T_Comp[kind] : A_Comp[kind]) != NULL;
 }
@ -643,7 +619,7 @@ void Compiler::Reset()
    LoadStorePatches.clear();
 }

-bool Compiler::IsJITFault(const u8* addr)
+bool Compiler::IsJITFault(u8* addr)
 {
    return (u64)addr >= (u64)ResetStart && (u64)addr < (u64)ResetStart + CodeMemSize;
 }
@ -658,45 +634,22 @@ void Compiler::Comp_SpecialBranchBehaviour(bool taken)
    {
        RegCache.PrepareExit();

-        if (ConstantCycles)
-            ADD(32, MDisp(RCPU, offsetof(ARM, Cycles)), Imm32(ConstantCycles));
-        ABI_TailCall(ARM_Ret);
+        ADD(32, MDisp(RCPU, offsetof(ARM, Cycles)), Imm32(ConstantCycles));
+        JMP((u8*)&ARM_Ret, true);
    }
 }

-#ifdef JIT_PROFILING_ENABLED
-void Compiler::CreateMethod(const char* namefmt, void* start, ...)
-{
-    if (iJIT_IsProfilingActive())
-    {
-        va_list args;
-        va_start(args, start);
-        char name[64];
-        vsprintf(name, namefmt, args);
-        va_end(args);
-
-        iJIT_Method_Load method = {0};
-        method.method_id = iJIT_GetNewMethodID();
-        method.method_name = name;
-        method.method_load_address = start;
-        method.method_size = GetWritableCodePtr() - (u8*)start;
-
-        iJIT_NotifyEvent(iJVM_EVENT_TYPE_METHOD_LOAD_FINISHED, (void*)&method);
-    }
-}
-#endif
-
-JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[], int instrsCount, bool hasMemoryInstr)
+JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[], int instrsCount)
 {
    if (NearSize - (GetCodePtr() - NearStart) < 1024 * 32) // guess...
    {
-        Log(LogLevel::Debug, "near reset\n");
-        NDS.JIT.ResetBlockCache();
+        printf("near reset\n");
+        ResetBlockCache();
    }
    if (FarSize - (FarCode - FarStart) < 1024 * 32) // guess...
    {
-        Log(LogLevel::Debug, "far reset\n");
-        NDS.JIT.ResetBlockCache();
+        printf("far reset\n");
+        ResetBlockCache();
    }

    ConstantCycles = 0;
@ -750,9 +703,7 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]
                ABI_CallFunction(InterpretTHUMB[CurInstr.Info.Kind]);
            }
            else
-            {
                (this->*comp)();
-            }
        }
        else
        {
@ -773,7 +724,7 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]
            }
            else
            {
-                IrregularCycles = comp == NULL;
+                IrregularCycles = false;

                FixupBranch skipExecute;
                if (cond < 0xE)
@ -786,9 +737,7 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]
                    ABI_CallFunction(InterpretARM[CurInstr.Info.Kind]);
                }
                else
-                {
                    (this->*comp)();
-                }

                Comp_SpecialBranchBehaviour(true);

@ -806,11 +755,9 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]
                        SetJumpTarget(skipFailed);
                    }
                    else
-                    {
                        SetJumpTarget(skipExecute);
-                    }
                }
-
+                
            }
        }

@ -820,13 +767,8 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]

    RegCache.Flush();

-    if (ConstantCycles)
-        ADD(32, MDisp(RCPU, offsetof(ARM, Cycles)), Imm32(ConstantCycles));
-    ABI_TailCall(ARM_Ret);
-
-#ifdef JIT_PROFILING_ENABLED
-    CreateMethod("JIT_Block_%d_%d_%08X", (void*)res, Num, Thumb, instrs[0].Addr);
-#endif
+    ADD(32, MDisp(RCPU, offsetof(ARM, Cycles)), Imm32(ConstantCycles));
+    JMP((u8*)ARM_Ret, true);

    /*FILE* codeout = fopen("codeout", "a");
    fprintf(codeout, "beginning block argargarg__ %x!!!", instrs[0].Addr);
@ -840,7 +782,7 @@ JitBlockEntry Compiler::CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[]
 void Compiler::Comp_AddCycles_C(bool forceNonConstant)
 {
    s32 cycles = Num ?
-        NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 1 : 3]
+        NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 1 : 3]
        : ((R15 & 0x2) ? 0 : CurInstr.CodeCycles);

    if ((!Thumb && CurInstr.Cond() < 0xE) || forceNonConstant)
@ -852,7 +794,7 @@ void Compiler::Comp_AddCycles_C(bool forceNonConstant)
 void Compiler::Comp_AddCycles_CI(u32 i)
 {
    s32 cycles = (Num ?
-        NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2]
+        NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2]
        : ((R15 & 0x2) ? 0 : CurInstr.CodeCycles)) + i;

    if (!Thumb && CurInstr.Cond() < 0xE)
@ -864,9 +806,9 @@ void Compiler::Comp_AddCycles_CI(u32 i)
 void Compiler::Comp_AddCycles_CI(Gen::X64Reg i, int add)
 {
    s32 cycles = Num ?
-        NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2]
+        NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2]
        : ((R15 & 0x2) ? 0 : CurInstr.CodeCycles);
-
+    
    if (!Thumb && CurInstr.Cond() < 0xE)
    {
        LEA(32, RSCRATCH, MDisp(i, add + cycles));
@ -889,7 +831,7 @@ void Compiler::Comp_AddCycles_CDI()

        s32 cycles;

-        s32 numC = NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2];
+        s32 numC = NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2];
        s32 numD = CurInstr.DataCycles;

        if ((CurInstr.DataRegion >> 24) == 0x02) // mainRAM
@ -911,7 +853,7 @@ void Compiler::Comp_AddCycles_CDI()
        {
            cycles = numC + numD + 1;
        }
-
+        
        if (!Thumb && CurInstr.Cond() < 0xE)
            ADD(32, MDisp(RCPU, offsetof(ARM, Cycles)), Imm8(cycles));
        else
@ -934,7 +876,7 @@ void Compiler::Comp_AddCycles_CD()
    }
    else
    {
-        s32 numC = NDS.ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2];
+        s32 numC = NDS::ARM7MemTimings[CurInstr.CodeCycles][Thumb ? 0 : 2];
        s32 numD = CurInstr.DataCycles;

        if ((CurInstr.DataRegion >> 4) == 0x02)
--- a/src/ARMJIT_x64/ARMJIT_Compiler.h
+++ b/src/ARMJIT_x64/ARMJIT_Compiler.h
@ -1,43 +1,17 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #ifndef ARMJIT_X64_COMPILER_H
 #define ARMJIT_X64_COMPILER_H

-#if defined(JIT_ENABLED) && defined(__x86_64__)
-
 #include "../dolphin/x64Emitter.h"

+#include "../ARMJIT.h"
 #include "../ARMJIT_Internal.h"
 #include "../ARMJIT_RegisterCache.h"

-#ifdef JIT_PROFILING_ENABLED
-#include <jitprofiling.h>
-#endif
-
 #include <unordered_map>

-
-namespace melonDS
+namespace ARMJIT
 {
-class ARMJIT;
-class ARMJIT_Memory;
-class NDS;
+
 const Gen::X64Reg RCPU = Gen::RBP;
 const Gen::X64Reg RCPSR = Gen::R15;

@ -83,17 +57,16 @@ struct Op2
 class Compiler : public Gen::XEmitter
 {
 public:
-    explicit Compiler(melonDS::NDS& nds);
-    ~Compiler();
+    Compiler();

    void Reset();

-    JitBlockEntry CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[], int instrsCount, bool hasMemoryInstr);
+    JitBlockEntry CompileBlock(ARM* cpu, bool thumb, FetchedInstr instrs[], int instrsCount);

    void LoadReg(int reg, Gen::X64Reg nativeReg);
    void SaveReg(int reg, Gen::X64Reg nativeReg);

-    bool CanCompile(bool thumb, u16 kind) const;
+    bool CanCompile(bool thumb, u16 kind);

    typedef void (Compiler::*CompileFunc)();

@ -126,7 +99,7 @@ public:
    void A_Comp_Mul_Long();

    void A_Comp_CLZ();
-
+    
    void A_Comp_MemWB();
    void A_Comp_MemHalf();
    void A_Comp_LDM_STM();
@ -172,10 +145,10 @@ public:
        memop_SubtractOffset = 1 << 4
    };
    void Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flags);
-    s32 Comp_MemAccessBlock(int rn, Common::BitSet16 regs, bool store, bool preinc, bool decrement, bool usermode, bool skipLoadingRn);
+    s32 Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc, bool decrement, bool usermode);
    bool Comp_MemLoadLiteral(int size, bool signExtend, int rd, u32 addr);

-    void Comp_ArithTriOp(void (Compiler::*op)(int, const Gen::OpArg&, const Gen::OpArg&),
+    void Comp_ArithTriOp(void (Compiler::*op)(int, const Gen::OpArg&, const Gen::OpArg&), 
        Gen::OpArg rd, Gen::OpArg rn, Gen::OpArg op2, bool carryUsed, int opFlags);
    void Comp_ArithTriOpReverse(void (Compiler::*op)(int, const Gen::OpArg&, const Gen::OpArg&),
        Gen::OpArg rd, Gen::OpArg rn, Gen::OpArg op2, bool carryUsed, int opFlags);
@ -201,8 +174,8 @@ public:

    Gen::FixupBranch CheckCondition(u32 cond);

-    void PushRegs(bool saveHiRegs, bool saveRegsToBeChanged, bool allowUnload = true);
-    void PopRegs(bool saveHiRegs, bool saveRegsToBeChanged);
+    void PushRegs(bool saveHiRegs);
+    void PopRegs(bool saveHiRegs);

    Gen::OpArg MapReg(int reg)
    {
@ -235,55 +208,48 @@ public:
        SetCodePtr(FarCode);
    }

-    bool IsJITFault(const u8* addr);
+    bool IsJITFault(u8* addr);

    u8* RewriteMemAccess(u8* pc);

-#ifdef JIT_PROFILING_ENABLED
-    void CreateMethod(const char* namefmt, void* start, ...);
-#endif
+    u8* FarCode;
+    u8* NearCode;
+    u32 FarSize;
+    u32 NearSize;

-    melonDS::NDS& NDS;
-    u8* FarCode {};
-    u8* NearCode {};
-    u32 FarSize {};
-    u32 NearSize {};
+    u8* NearStart;
+    u8* FarStart;

-    u8* NearStart {};
-    u8* FarStart {};
+    void* PatchedStoreFuncs[2][2][3][16];
+    void* PatchedLoadFuncs[2][2][3][2][16];

-    void* PatchedStoreFuncs[2][2][3][16] {};
-    void* PatchedLoadFuncs[2][2][3][2][16] {};
+    std::unordered_map<u8*, LoadStorePatch> LoadStorePatches;

-    std::unordered_map<u8*, LoadStorePatch> LoadStorePatches {};
+    u8* ResetStart;
+    u32 CodeMemSize;

-    u8* CodeMemBase;
-    u8* ResetStart {};
-    u32 CodeMemSize {};
+    bool Exit;
+    bool IrregularCycles;

-    bool Exit {};
-    bool IrregularCycles {};
-
-    void* ReadBanked {};
-    void* WriteBanked {};
+    void* ReadBanked;
+    void* WriteBanked;

    bool CPSRDirty = false;

-    FetchedInstr CurInstr {};
+    FetchedInstr CurInstr;

-    RegisterCache<Compiler, Gen::X64Reg> RegCache {};
+    RegisterCache<Compiler, Gen::X64Reg> RegCache;

-    bool Thumb {};
-    u32 Num {};
-    u32 R15 {};
-    u32 CodeRegion {};
+    bool Thumb;
+    u32 Num;
+    u32 R15;
+    u32 CodeRegion;

-    u32 ConstantCycles {};
+    u32 ConstantCycles;

-    ARM* CurCPU {};
+    ARM* CurCPU;
 };

 }
-#endif

 #endif
--- a/src/ARMJIT_x64/ARMJIT_GenOffsets.cpp
+++ b/src/ARMJIT_x64/ARMJIT_GenOffsets.cpp
@ -1,23 +1,5 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "../ARM.h"
-using namespace melonDS;
+
 int main(int argc, char* argv[])
 {
    FILE* f = fopen("ARMJIT_Offsets.h", "w");
--- a/src/ARMJIT_x64/ARMJIT_Linkage.S
+++ b/src/ARMJIT_x64/ARMJIT_Linkage.S
@ -1,21 +1,3 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 .intel_syntax noprefix

 #include "ARMJIT_Offsets.h"
@ -104,8 +86,3 @@ ARM_Ret:
 #endif

    ret
-
-#if !defined(__APPLE__) && !defined(WIN64)
-.section .note.GNU-stack,"",@progbits
-#endif
-
--- a/src/ARMJIT_x64/ARMJIT_LoadStore.cpp
+++ b/src/ARMJIT_x64/ARMJIT_LoadStore.cpp
@ -1,28 +1,10 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARMJIT_Compiler.h"
-#include "../ARMJIT.h"
-#include "../NDS.h"
+
+#include "../Config.h"

 using namespace Gen;

-namespace melonDS
+namespace ARMJIT
 {

 template <typename T>
@ -53,7 +35,7 @@ u8* Compiler::RewriteMemAccess(u8* pc)
        return pc + (ptrdiff_t)patch.Offset;
    }

-    Log(LogLevel::Error, "this is a JIT bug %sx\n", pc);
+    printf("this is a JIT bug %llx\n", pc);
    abort();
 }

@ -69,11 +51,12 @@ u8* Compiler::RewriteMemAccess(u8* pc)

 bool Compiler::Comp_MemLoadLiteral(int size, bool signExtend, int rd, u32 addr)
 {
-    u32 localAddr = NDS.JIT.LocaliseCodeAddress(Num, addr);
+    u32 localAddr = LocaliseCodeAddress(Num, addr);

-    int invalidLiteralIdx = NDS.JIT.InvalidLiterals.Find(localAddr);
+    int invalidLiteralIdx = InvalidLiterals.Find(localAddr);
    if (invalidLiteralIdx != -1)
    {
+        InvalidLiterals.Remove(invalidLiteralIdx);
        return false;
    }

@ -86,7 +69,7 @@ bool Compiler::Comp_MemLoadLiteral(int size, bool signExtend, int rd, u32 addr)
    if (size == 32)
    {
        CurCPU->DataRead32(addr & ~0x3, &val);
-        val = melonDS::ROR(val, (addr & 0x3) << 3);
+        val = ::ROR(val, (addr & 0x3) << 3);
    }
    else if (size == 16)
    {
@ -106,7 +89,7 @@ bool Compiler::Comp_MemLoadLiteral(int size, bool signExtend, int rd, u32 addr)

    if (Thumb || CurInstr.Cond() == 0xE)
        RegCache.PutLiteral(rd, val);
-
+    
    return true;
 }

@ -119,10 +102,10 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
    if (size == 16)
        addressMask = ~1;

-    if (NDS.JIT.LiteralOptimizationsEnabled() && rn == 15 && rd != 15 && op2.IsImm && !(flags & (memop_Post|memop_Store|memop_Writeback)))
+    if (Config::JIT_LiteralOptimisations && rn == 15 && rd != 15 && op2.IsImm && !(flags & (memop_Post|memop_Store|memop_Writeback)))
    {
        u32 addr = R15 + op2.Imm * ((flags & memop_SubtractOffset) ? -1 : 1);
-
+        
        if (Comp_MemLoadLiteral(size, flags & memop_SignExtend, rd, addr))
            return;
    }
@ -136,7 +119,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
        Comp_AddCycles_CDI();
    }

-    bool addrIsStatic = NDS.JIT.LiteralOptimizationsEnabled()
+    bool addrIsStatic = Config::JIT_LiteralOptimisations
        && RegCache.IsLiteral(rn) && op2.IsImm && !(flags & (memop_Writeback|memop_Post));
    u32 staticAddress;
    if (addrIsStatic)
@ -147,12 +130,6 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
    if (Thumb && rn == 15)
        rnMapped = Imm32(R15 & ~0x2);

-    if (flags & memop_Store && flags & (memop_Post|memop_Writeback) && rd == rn)
-    {
-        MOV(32, R(RSCRATCH4), rdMapped);
-        rdMapped = R(RSCRATCH4);
-    }
-
    X64Reg finalAddr = RSCRATCH3;
    if (flags & memop_Post)
    {
@ -172,7 +149,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
        if (!(flags & memop_SubtractOffset) && rm.IsSimpleReg() && rnMapped.IsSimpleReg()
            && op2.Reg.Op == 0 && op2.Reg.Amount > 0 && op2.Reg.Amount <= 3)
        {
-            LEA(32, finalAddr,
+            LEA(32, finalAddr, 
                MComplex(rnMapped.GetSimpleReg(), rm.GetSimpleReg(), 1 << op2.Reg.Amount, 0));
        }
        else
@ -197,10 +174,10 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
        MOV(32, rnMapped, R(finalAddr));

    u32 expectedTarget = Num == 0
-        ? NDS.JIT.Memory.ClassifyAddress9(CurInstr.DataRegion)
-        : NDS.JIT.Memory.ClassifyAddress7(CurInstr.DataRegion);
+        ? ARMJIT_Memory::ClassifyAddress9(CurInstr.DataRegion)
+        : ARMJIT_Memory::ClassifyAddress7(CurInstr.DataRegion);

-    if (NDS.JIT.FastMemoryEnabled() && ((!Thumb && CurInstr.Cond() != 0xE) || NDS.JIT.Memory.IsFastmemCompatible(expectedTarget)))
+    if (Config::JIT_FastMemory && ((!Thumb && CurInstr.Cond() != 0xE) || ARMJIT_Memory::IsFastmemCompatible(expectedTarget)))
    {
        if (rdMapped.IsImm())
        {
@ -213,12 +190,12 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag

        assert(rdMapped.GetSimpleReg() >= 0 && rdMapped.GetSimpleReg() < 16);
        patch.PatchFunc = flags & memop_Store
-            ? PatchedStoreFuncs[NDS.ConsoleType][Num][__builtin_ctz(size) - 3][rdMapped.GetSimpleReg()]
-            : PatchedLoadFuncs[NDS.ConsoleType][Num][__builtin_ctz(size) - 3][!!(flags & memop_SignExtend)][rdMapped.GetSimpleReg()];
+            ? PatchedStoreFuncs[NDS::ConsoleType][Num][__builtin_ctz(size) - 3][rdMapped.GetSimpleReg()]
+            : PatchedLoadFuncs[NDS::ConsoleType][Num][__builtin_ctz(size) - 3][!!(flags & memop_SignExtend)][rdMapped.GetSimpleReg()];

        assert(patch.PatchFunc != NULL);

-        MOV(64, R(RSCRATCH), ImmPtr(Num == 0 ? NDS.JIT.Memory.FastMem9Start : NDS.JIT.Memory.FastMem7Start));
+        MOV(64, R(RSCRATCH), ImmPtr(Num == 0 ? ARMJIT_Memory::FastMem9Start : ARMJIT_Memory::FastMem7Start));

        X64Reg maskedAddr = RSCRATCH3;
        if (size > 8)
@ -265,11 +242,11 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
    }
    else
    {
-        PushRegs(false, false);
+        PushRegs(false);

        void* func = NULL;
        if (addrIsStatic)
-            func = NDS.JIT.Memory.GetFuncForAddr(CurCPU, staticAddress, flags & memop_Store, size);
+            func = ARMJIT_Memory::GetFuncForAddr(CurCPU, staticAddress, flags & memop_Store, size);

        if (func)
        {
@ -282,7 +259,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag

            ABI_CallFunction((void (*)())func);

-            PopRegs(false, false);
+            PopRegs(false);

            if (!(flags & memop_Store))
            {
@ -305,35 +282,33 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
        {
            if (Num == 0)
            {
-                // on Windows param 3 is R8 which is also scratch 4 which can be used for rd
-                if (flags & memop_Store)
-                    MOV(32, R(ABI_PARAM3), rdMapped);
-
                MOV(64, R(ABI_PARAM2), R(RCPU));
                if (ABI_PARAM1 != RSCRATCH3)
                    MOV(32, R(ABI_PARAM1), R(RSCRATCH3));
                if (flags & memop_Store)
                {
-                    switch (size | NDS.ConsoleType)
+                    MOV(32, R(ABI_PARAM3), rdMapped);
+
+                    switch (size | NDS::ConsoleType)
                    {
-                    case 32: ABI_CallFunction(SlowWrite9<u32, 0>); break;
-                    case 16: ABI_CallFunction(SlowWrite9<u16, 0>); break;
-                    case 8: ABI_CallFunction(&SlowWrite9<u8, 0>); break;
-                    case 33: ABI_CallFunction(&SlowWrite9<u32, 1>); break;
-                    case 17: ABI_CallFunction(&SlowWrite9<u16, 1>); break;
-                    case 9: ABI_CallFunction(&SlowWrite9<u8, 1>); break;
+                    case 32: CALL((void*)&SlowWrite9<u32, 0>); break;
+                    case 16: CALL((void*)&SlowWrite9<u16, 0>); break;
+                    case 8: CALL((void*)&SlowWrite9<u8, 0>); break;
+                    case 33: CALL((void*)&SlowWrite9<u32, 1>); break;
+                    case 17: CALL((void*)&SlowWrite9<u16, 1>); break;
+                    case 9: CALL((void*)&SlowWrite9<u8, 1>); break;
                    }
                }
                else
                {
-                    switch (size | NDS.ConsoleType)
+                    switch (size | NDS::ConsoleType)
                    {
-                    case 32: ABI_CallFunction(&SlowRead9<u32, 0>); break;
-                    case 16: ABI_CallFunction(&SlowRead9<u16, 0>); break;
-                    case 8: ABI_CallFunction(&SlowRead9<u8, 0>); break;
-                    case 33: ABI_CallFunction(&SlowRead9<u32, 1>); break;
-                    case 17: ABI_CallFunction(&SlowRead9<u16, 1>); break;
-                    case 9: ABI_CallFunction(&SlowRead9<u8, 1>); break;
+                    case 32: CALL((void*)&SlowRead9<u32, 0>); break;
+                    case 16: CALL((void*)&SlowRead9<u16, 0>); break;
+                    case 8: CALL((void*)&SlowRead9<u8, 0>); break;
+                    case 33: CALL((void*)&SlowRead9<u32, 1>); break;
+                    case 17: CALL((void*)&SlowRead9<u16, 1>); break;
+                    case 9: CALL((void*)&SlowRead9<u8, 1>); break;
                    }
                }
            }
@ -345,32 +320,32 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
                {
                    MOV(32, R(ABI_PARAM2), rdMapped);

-                    switch (size | NDS.ConsoleType)
+                    switch (size | NDS::ConsoleType)
                    {
-                    case 32: ABI_CallFunction(&SlowWrite7<u32, 0>); break;
-                    case 16: ABI_CallFunction(&SlowWrite7<u16, 0>); break;
-                    case 8: ABI_CallFunction(&SlowWrite7<u8, 0>); break;
-                    case 33: ABI_CallFunction(&SlowWrite7<u32, 1>); break;
-                    case 17: ABI_CallFunction(&SlowWrite7<u16, 1>); break;
-                    case 9: ABI_CallFunction(&SlowWrite7<u8, 1>); break;
+                    case 32: CALL((void*)&SlowWrite7<u32, 0>); break;
+                    case 16: CALL((void*)&SlowWrite7<u16, 0>); break;
+                    case 8: CALL((void*)&SlowWrite7<u8, 0>); break;
+                    case 33: CALL((void*)&SlowWrite7<u32, 1>); break;
+                    case 17: CALL((void*)&SlowWrite7<u16, 1>); break;
+                    case 9: CALL((void*)&SlowWrite7<u8, 1>); break;
                    }
                }
                else
                {
-                    switch (size | NDS.ConsoleType)
+                    switch (size | NDS::ConsoleType)
                    {
-                    case 32: ABI_CallFunction(&SlowRead7<u32, 0>); break;
-                    case 16: ABI_CallFunction(&SlowRead7<u16, 0>); break;
-                    case 8: ABI_CallFunction(&SlowRead7<u8, 0>); break;
-                    case 33: ABI_CallFunction(&SlowRead7<u32, 1>); break;
-                    case 17: ABI_CallFunction(&SlowRead7<u16, 1>); break;
-                    case 9: ABI_CallFunction(&SlowRead7<u8, 1>); break;
+                    case 32: CALL((void*)&SlowRead7<u32, 0>); break;
+                    case 16: CALL((void*)&SlowRead7<u16, 0>); break;
+                    case 8: CALL((void*)&SlowRead7<u8, 0>); break;
+                    case 33: CALL((void*)&SlowRead7<u32, 1>); break;
+                    case 17: CALL((void*)&SlowRead7<u16, 1>); break;
+                    case 9: CALL((void*)&SlowRead7<u8, 1>); break;
                    }
                }
            }

-            PopRegs(false, false);
-
+            PopRegs(false);
+            
            if (!(flags & memop_Store))
            {
                if (flags & memop_SignExtend)
@ -384,7 +359,7 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
    if (!(flags & memop_Store) && rd == 15)
    {
        if (size < 32)
-            Log(LogLevel::Debug, "!!! LDR <32 bit PC %08X %x\n", R15, CurInstr.Instr);
+            printf("!!! LDR <32 bit PC %08X %x\n", R15, CurInstr.Instr);
        {
            if (Num == 1)
            {
@ -398,15 +373,14 @@ void Compiler::Comp_MemAccess(int rd, int rn, const Op2& op2, int size, int flag
    }
 }

-s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc, bool decrement, bool usermode, bool skipLoadingRn)
+s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc, bool decrement, bool usermode)
 {
    int regsCount = regs.Count();

    if (regsCount == 0)
        return 0; // actually not the right behaviour TODO: fix me

-    int firstReg = *regs.begin();
-    if (regsCount == 1 && !usermode && RegCache.LoadedRegs & (1 << firstReg) && !(firstReg == rn && skipLoadingRn))
+    if (regsCount == 1 && !usermode && RegCache.LoadedRegs & (1 << *regs.begin()))
    {
        int flags = 0;
        if (store)
@ -415,7 +389,7 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
            flags |= memop_SubtractOffset;
        Op2 offset = preinc ? Op2(4) : Op2(0);

-        Comp_MemAccess(firstReg, rn, offset, 32, flags);
+        Comp_MemAccess(*regs.begin(), rn, offset, 32, flags);

        return decrement ? -4 : 4;
    }
@ -423,16 +397,16 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
    s32 offset = (regsCount * 4) * (decrement ? -1 : 1);

    int expectedTarget = Num == 0
-        ? NDS.JIT.Memory.ClassifyAddress9(CurInstr.DataRegion)
-        : NDS.JIT.Memory.ClassifyAddress7(CurInstr.DataRegion);
+        ? ARMJIT_Memory::ClassifyAddress9(CurInstr.DataRegion)
+        : ARMJIT_Memory::ClassifyAddress7(CurInstr.DataRegion);

    if (!store)
        Comp_AddCycles_CDI();
    else
        Comp_AddCycles_CD();

-    bool compileFastPath = NDS.JIT.FastMemoryEnabled()
-        && !usermode && (CurInstr.Cond() < 0xE || NDS.JIT.Memory.IsFastmemCompatible(expectedTarget));
+    bool compileFastPath = Config::JIT_FastMemory
+        && !usermode && (CurInstr.Cond() < 0xE || ARMJIT_Memory::IsFastmemCompatible(expectedTarget));

    // we need to make sure that the stack stays aligned to 16 bytes
 #ifdef _WIN32
@ -455,7 +429,7 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
        u8* fastPathStart = GetWritableCodePtr();
        u8* loadStoreAddr[16];

-        MOV(64, R(RSCRATCH2), ImmPtr(Num == 0 ? NDS.JIT.Memory.FastMem9Start : NDS.JIT.Memory.FastMem7Start));
+        MOV(64, R(RSCRATCH2), ImmPtr(Num == 0 ? ARMJIT_Memory::FastMem9Start : ARMJIT_Memory::FastMem7Start));
        ADD(64, R(RSCRATCH2), R(RSCRATCH4));

        u32 offset = 0;
@ -482,10 +456,7 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
            {
                if (RegCache.LoadedRegs & (1 << reg))
                {
-                    if (!(reg == rn && skipLoadingRn))
-                        MOV(32, MapReg(reg), mem);
-                    else
-                        MOV(32, R(RSCRATCH), mem); // just touch the memory
+                    MOV(32, MapReg(reg), mem);
                }
                else
                {
@ -511,7 +482,7 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc

    if (!store)
    {
-        PushRegs(false, false, !compileFastPath);
+        PushRegs(false);

        MOV(32, R(ABI_PARAM1), R(RSCRATCH4));
        MOV(32, R(ABI_PARAM3), Imm32(regsCount));
@ -524,15 +495,15 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
        if (Num == 0)
            MOV(64, R(ABI_PARAM4), R(RCPU));

-        switch (Num * 2 | NDS.ConsoleType)
+        switch (Num * 2 | NDS::ConsoleType)
        {
-        case 0: ABI_CallFunction(&SlowBlockTransfer9<false, 0>); break;
-        case 1: ABI_CallFunction(&SlowBlockTransfer9<false, 1>); break;
-        case 2: ABI_CallFunction(&SlowBlockTransfer7<false, 0>); break;
-        case 3: ABI_CallFunction(&SlowBlockTransfer7<false, 1>); break;
+        case 0: CALL((void*)&SlowBlockTransfer9<false, 0>); break;
+        case 1: CALL((void*)&SlowBlockTransfer9<false, 1>); break;
+        case 2: CALL((void*)&SlowBlockTransfer7<false, 0>); break;
+        case 3: CALL((void*)&SlowBlockTransfer7<false, 1>); break;
        }

-        PopRegs(false, false);
+        PopRegs(false);

        if (allocOffset)
            ADD(64, R(RSP), Imm8(allocOffset));
@ -551,15 +522,12 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
                MOV(32, R(RSCRATCH2), Imm32(reg - 8));
                POP(RSCRATCH3);
                CALL(WriteBanked);
-                if (!(reg == rn && skipLoadingRn))
-                {
-                    FixupBranch sucessfulWritten = J_CC(CC_NC);
-                    if (RegCache.LoadedRegs & (1 << reg))
-                            MOV(32, R(RegCache.Mapping[reg]), R(RSCRATCH3));
-                    else
-                        SaveReg(reg, RSCRATCH3);
-                    SetJumpTarget(sucessfulWritten);
-                }
+                FixupBranch sucessfulWritten = J_CC(CC_NC);
+                if (RegCache.LoadedRegs & (1 << reg))
+                    MOV(32, R(RegCache.Mapping[reg]), R(RSCRATCH3));
+                else
+                    SaveReg(reg, RSCRATCH3);
+                SetJumpTarget(sucessfulWritten);
            }
            else if (!(RegCache.LoadedRegs & (1 << reg)))
            {
@ -568,10 +536,6 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
                POP(RSCRATCH);
                SaveReg(reg, RSCRATCH);
            }
-            else if (reg == rn && skipLoadingRn)
-            {
-                ADD(64, R(RSP), Imm8(8));
-            }
            else
            {
                POP(MapReg(reg).GetSimpleReg());
@ -616,29 +580,29 @@ s32 Compiler::Comp_MemAccessBlock(int rn, BitSet16 regs, bool store, bool preinc
        if (allocOffset)
            SUB(64, R(RSP), Imm8(allocOffset));

-        PushRegs(false, false, !compileFastPath);
+        PushRegs(false);

        MOV(32, R(ABI_PARAM1), R(RSCRATCH4));
        if (allocOffset)
            LEA(64, ABI_PARAM2, MDisp(RSP, allocOffset));
        else
            MOV(64, R(ABI_PARAM2), R(RSP));
-
+        
        MOV(32, R(ABI_PARAM3), Imm32(regsCount));
        if (Num == 0)
            MOV(64, R(ABI_PARAM4), R(RCPU));

-        switch (Num * 2 | NDS.ConsoleType)
+        switch (Num * 2 | NDS::ConsoleType)
        {
-        case 0: ABI_CallFunction(&SlowBlockTransfer9<true, 0>); break;
-        case 1: ABI_CallFunction(&SlowBlockTransfer9<true, 1>); break;
-        case 2: ABI_CallFunction(&SlowBlockTransfer7<true, 0>); break;
-        case 3: ABI_CallFunction(&SlowBlockTransfer7<true, 1>); break;
+        case 0: CALL((void*)&SlowBlockTransfer9<true, 0>); break;
+        case 1: CALL((void*)&SlowBlockTransfer9<true, 1>); break;
+        case 2: CALL((void*)&SlowBlockTransfer7<true, 0>); break;
+        case 3: CALL((void*)&SlowBlockTransfer7<true, 1>); break;
        }

        ADD(64, R(RSP), stackAlloc <= INT8_MAX ? Imm8(stackAlloc) : Imm32(stackAlloc));
-
-        PopRegs(false, false);
+    
+        PopRegs(false);
    }

    if (compileFastPath)
@ -668,7 +632,7 @@ void Compiler::A_Comp_MemWB()
    bool load = CurInstr.Instr & (1 << 20);
    bool byte = CurInstr.Instr & (1 << 22);
    int size = byte ? 8 : 32;
-
+    
    int flags = 0;
    if (!load)
        flags |= memop_Store;
@ -742,7 +706,7 @@ void Compiler::T_Comp_MemReg()
    bool load = op & 0x2;
    bool byte = op & 0x1;

-    Comp_MemAccess(CurInstr.T_Reg(0), CurInstr.T_Reg(3), Op2(CurInstr.T_Reg(6), 0, 0),
+    Comp_MemAccess(CurInstr.T_Reg(0), CurInstr.T_Reg(3), Op2(CurInstr.T_Reg(6), 0, 0), 
        byte ? 8 : 32, load ? 0 : memop_Store);
 }

@ -758,14 +722,14 @@ void Compiler::A_Comp_LDM_STM()

    OpArg rn = MapReg(CurInstr.A_Reg(16));

+    s32 offset = Comp_MemAccessBlock(CurInstr.A_Reg(16), regs, !load, pre, !add, usermode);
+
    if (load && writeback && regs[CurInstr.A_Reg(16)])
        writeback = Num == 0
-            && (!(regs & ~BitSet16(1 << CurInstr.A_Reg(16)))) || (regs & ~BitSet16((2 << CurInstr.A_Reg(16)) - 1));
-
-    s32 offset = Comp_MemAccessBlock(CurInstr.A_Reg(16), regs, !load, pre, !add, usermode, load && writeback);
-
-    if (writeback && offset)
-        ADD(32, rn, Imm32(offset));
+            ? (!(regs & ~BitSet16(1 << CurInstr.A_Reg(16)))) || (regs & ~BitSet16((2 << CurInstr.A_Reg(16)) - 1))
+            : false;
+    if (writeback)
+        ADD(32, rn, offset >= INT8_MIN && offset < INT8_MAX ? Imm8(offset) : Imm32(offset));
 }

 void Compiler::T_Comp_MemImm()
@ -809,7 +773,7 @@ void Compiler::T_Comp_LoadPCRel()
 {
    u32 offset = (CurInstr.Instr & 0xFF) << 2;
    u32 addr = (R15 & ~0x2) + offset;
-    if (!NDS.JIT.LiteralOptimizationsEnabled() || !Comp_MemLoadLiteral(32, false, CurInstr.T_Reg(8), addr))
+    if (!Config::JIT_LiteralOptimisations || !Comp_MemLoadLiteral(32, false, CurInstr.T_Reg(8), addr))
        Comp_MemAccess(CurInstr.T_Reg(8), 15, Op2(offset), 32, 0);
 }

@ -835,10 +799,9 @@ void Compiler::T_Comp_PUSH_POP()
    }

    OpArg sp = MapReg(13);
-    s32 offset = Comp_MemAccessBlock(13, regs, !load, !load, !load, false, false);
+    s32 offset = Comp_MemAccessBlock(13, regs, !load, !load, !load, false);

-    if (offset)
-        ADD(32, sp, Imm8(offset)); // offset will be always be in range since PUSH accesses 9 regs max
+    ADD(32, sp, Imm8(offset)); // offset will be always be in range since PUSH accesses 9 regs max
 }

 void Compiler::T_Comp_LDMIA_STMIA()
@ -847,11 +810,9 @@ void Compiler::T_Comp_LDMIA_STMIA()
    OpArg rb = MapReg(CurInstr.T_Reg(8));
    bool load = CurInstr.Instr & (1 << 11);

-    bool writeback = !load || !regs[CurInstr.T_Reg(8)];
+    s32 offset = Comp_MemAccessBlock(CurInstr.T_Reg(8), regs, !load, false, false, false);

-    s32 offset = Comp_MemAccessBlock(CurInstr.T_Reg(8), regs, !load, false, false, false, load && writeback);
-
-    if (writeback && offset)
+    if (!load || !regs[CurInstr.T_Reg(8)])
        ADD(32, rb, Imm8(offset));
 }

--- a/src/ARMJIT_x64/ARMJIT_Offsets.h
+++ b/src/ARMJIT_x64/ARMJIT_Offsets.h
@ -1,21 +1,3 @@
-/*
-    Copyright 2016-2024 melonDS team, RSDuck
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #define ARM_CPSR_offset 0x64
 #define ARM_Cycles_offset 0xc
 #define ARM_StopExecution_offset 0x10
--- a/src/ARM_InstrInfo.cpp
+++ b/src/ARM_InstrInfo.cpp
@ -1,28 +1,10 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #include "ARM_InstrInfo.h"

 #include <stdio.h>

-#include "ARMJIT.h"
+#include "Config.h"

-namespace melonDS::ARMInstrInfo
+namespace ARMInstrInfo
 {

 #define ak(x) ((x) << 23)
@ -230,7 +212,7 @@ enum {
    T_SetMaybeC     = 1 << 17,
    T_ReadC         = 1 << 18,
    T_SetC          = 1 << 19,
-
+    
    T_WriteMem      = 1 << 20,
    T_LoadMem       = 1 << 21,
 };
@ -315,7 +297,7 @@ const u32 T_SVC = T_BranchAlways | T_WriteR14 | tk(tk_SVC);
 #include "ARM_InstrTable.h"
 #undef INSTRFUNC_PROTO

-Info Decode(bool thumb, u32 num, u32 instr, bool literaloptimizations)
+Info Decode(bool thumb, u32 num, u32 instr)
 {
    const u8 FlagsReadPerCond[7] = {
        flag_Z,
@ -345,7 +327,7 @@ Info Decode(bool thumb, u32 num, u32 instr, bool literaloptimizations)
            res.DstRegs |= 1 << (instr & 0x7);
        if (data & T_Write8)
            res.DstRegs |= 1 << ((instr >> 8) & 0x7);
-
+        
        if (data & T_ReadHi0)
            res.SrcRegs |= 1 << ((instr & 0x7) | ((instr >> 4) & 0x8));
        if (data & T_ReadHi3)
@ -381,12 +363,12 @@ Info Decode(bool thumb, u32 num, u32 instr, bool literaloptimizations)

        if (data & T_WriteMem)
            res.SpecialKind = special_WriteMem;
-
+        
        if (data & T_LoadMem)
        {
            if (res.Kind == tk_LDR_PCREL)
            {
-                if (!literaloptimizations)
+                if (!Config::JIT_LiteralOptimisations)
                    res.SrcRegs |= 1 << 15;
                res.SpecialKind = special_LoadLiteral;
            }
@ -471,18 +453,18 @@ Info Decode(bool thumb, u32 num, u32 instr, bool literaloptimizations)
            res.SrcRegs |= 1 << ((instr >> 8) & 0xF);
        if (data & A_Read12)
            res.SrcRegs |= 1 << ((instr >> 12) & 0xF);
-
+        
        if (data & A_Write12)
            res.DstRegs |= 1 << ((instr >> 12) & 0xF);
        if (data & A_Write16)
            res.DstRegs |= 1 << ((instr >> 16) & 0xF);
-
+        
        if (data & A_MemWriteback && instr & (1 << 21))
            res.DstRegs |= 1 << ((instr >> 16) & 0xF);

        if (data & A_BranchAlways)
            res.DstRegs |= 1 << 15;
-
+        
        if (data & A_Read12Double)
        {
            res.SrcRegs |= 1 << ((instr >> 12) & 0xF);
@ -526,21 +508,16 @@ Info Decode(bool thumb, u32 num, u32 instr, bool literaloptimizations)
        if (data & A_LoadMem)
        {
            if (res.SrcRegs == (1 << 15))
-                res.SpecialKind = special_LoadLiteral;
+               res.SpecialKind = special_LoadLiteral;
            else
                res.SpecialKind = special_LoadMem;
        }
-
+        
        if (res.Kind == ak_LDM)
        {
            u16 set = (instr & 0xFFFF);
            res.NotStrictlyNeeded |= set & ~(res.SrcRegs|res.DstRegs|(1<<15));
            res.DstRegs |= set;
-            // when the instruction is executed not in usermode a banked register in memory will be written to
-            // but the unbanked register will still be allocated, so it is expected to carry the proper value
-            // thus it is a source register
-            if (instr & (1<<22))
-                res.SrcRegs |= set & 0x7F00;
        }
        if (res.Kind == ak_STM)
        {
--- a/src/ARM_InstrInfo.h
+++ b/src/ARM_InstrInfo.h
@ -1,27 +1,9 @@
-/*
-    Copyright 2016-2024 melonDS team
-
-    This file is part of melonDS.
-
-    melonDS is free software: you can redistribute it and/or modify it under
-    the terms of the GNU General Public License as published by the Free
-    Software Foundation, either version 3 of the License, or (at your option)
-    any later version.
-
-    melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
-    WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-    FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-    You should have received a copy of the GNU General Public License along
-    with melonDS. If not, see http://www.gnu.org/licenses/.
-*/
-
 #ifndef ARMINSTRINFO_H
 #define ARMINSTRINFO_H

 #include "types.h"

-namespace melonDS::ARMInstrInfo
+namespace ARMInstrInfo
 {

 // Instruction kinds, for faster dispatch
@ -222,7 +204,7 @@ enum
    tk_POP,
    tk_LDMIA,
    tk_STMIA,
-
+    
    tk_BCOND,
    tk_BX,
    tk_BLX_REG,
@ -274,8 +256,8 @@ struct Info
    }
 };

-Info Decode(bool thumb, u32 num, u32 instr, bool literaloptimizations);
+Info Decode(bool thumb, u32 num, u32 instr);

 }

-#endif
+#endif
--- a/src/ARM_InstrTable.h
+++ b/src/ARM_InstrTable.h
@ -1,5 +1,5 @@
 /*
-    Copyright 2016-2024 melonDS team
+    Copyright 2016-2020 Arisotura

    This file is part of melonDS.

--- a/Show More
+++ b/Show More
				`@ -1 +0,0 @@`
				`include(${CMAKE_CURRENT_LIST_DIR}/ECMFindModuleHelpers.cmake)`